Upload folder using huggingface_hub

.gitattributes

@@ -13582,3 +13582,51 @@ tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_l
 tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
 tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
 tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx16_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx16_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx16_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx16_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx16_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx16_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx16_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx16_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+OnZeta/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx16_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx16_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn101_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx16_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx16_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/rn50_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx16_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx16_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b16_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx16_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx16_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx16_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed1/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed2/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text
+Onzeta_2/tpt/pretrained_cocoop/vit_b32_ep50_16shots/nctx4_cscFalse_ctpend/seed3/prompt_learner/model.pth.tar-50 filter=lfs diff=lfs merge=lfs -text

OnZeta/11.log

@@ -0,0 +1,152 @@
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.23
+mean acc of onzeta is: 9.29
+mean acc of MAPLS is: 17.81
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.05)
+the beta is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.0)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.14
+mean acc of onzeta is: 9.65
+mean acc of MAPLS is: 17.93
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.0)
+the beta is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.14
+mean acc of onzeta is: 12.26
+mean acc of MAPLS is: 12.58
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5, entropy_weight=0.0)
+the beta is 0.8
+load pre-trained model
+load data

OnZeta/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

OnZeta/MAPLS/common.py

@@ -0,0 +1,146 @@
+# Common Utils in Target Label Shift Estimation
+import numpy as np
+from typing import List
+
+
+# Post hoc Label Shift Correction--------------------------------------#
+def lsc(probs: np.ndarray, w: List):
+    r"""
+    Implementation of Label Shift Compensation (LSC) with known target label distribution.
+    Given source domain P(Y_s=i) and P(Y_s=i|X_s=x), target domain P(Y_t=i),
+    estimate target predicted probability q(y|x) on test set.
+
+    Args:
+        probs:      Softmax probability P(Y_s=i|X_s=x) predicted by the classifier,
+                    for all samples in validation set (N x C).
+        w:          Ratio of Target over Source domain label distribution $ w = P(Y_t=i) / P(Y_s=i) $,
+                    Not necessarily normalized to 1.
+
+    Shapes:
+        * Input:
+            probs:     N x C   (No. of samples) x (No. of classes),
+            w:         C       (No. of classes),
+        * Output:
+            pc_probs:  N x C   (No. of samples) x (No. of classes)
+
+
+    For more information see original paper:
+    [2002] "Adjusting the Outputs of a Classifier to New a Priori Probabilities: A Simple Procedure"
+    """
+    assert len(w) == probs.shape[-1]
+    probs = probs.detach().numpy()
+    pc_probs = normalized(probs * w, axis=-1, order=1)
+
+    return pc_probs
+
+
+# Estimation of Source Label Distribution P(Y_s=i) or p(\hat{y}=c_i)----------#
+def get_py(probs: np.ndarray, cls_num_list: List[int] = None, mode='soft'):
+    r"""
+    Estimation of source label distribution (normalized)
+    Given source domain P(Y_s=i|X_s=x)=f(x) and No. of sample per-class,
+    estimate P(Y_s=i) or p(\hat{y}=c_i)
+
+    Args:
+        probs:          Softmax probability p(\hat{y}|x) predicted by classifier,
+                        over all samples on train set (N x C).
+        cls_num_list:   No. of samples in each class (C)
+        mode:           Method used to estimate p(\hat{y}=c_i),
+                        'soft' will estimate p(\hat{y}=c_i) \approx \sum^N_j p(\hat{y}=c_i|x_j),
+                        'hard' will estimate p(\hat{y}=c_i) \approx \sum^N_j \mathds{1}(\arg\max_c p(y=c|x_j)=c_i)
+                        'gt' will estimate p(\hat{y}=c_i) \approx P(Y_s=i), which is given by cls_num_list
+
+    Shapes:
+        * Input:
+            probs:          N x C   (No. of samples) x (No. of classes),
+            cls_num_list:   C       (No. of classes),
+        * Output:
+            py:             C       (No. of classes)
+
+    Examples:
+
+        >>> import numpy as np
+        >>> from numpy.linalg import norm
+        >>> class_num = 5; val_set_sample_num = 10
+        >>> prob = norm(np.random.normal(size=(val_set_sample_num, class_num)), ord=1, axis=-1)
+        >>> num_list = list(range(class_num))
+        >>> py = get_py(prob, num_list)
+    """
+    cls_num = probs.shape[-1]
+
+    if mode == "soft":
+        py = np.mean(probs, axis=tuple(range(len(np.shape(probs)) - 1)))
+    elif mode == "hard":
+        py = np.bincount(np.argmax(probs, axis=-1), minlength=cls_num)
+        py = py / py.sum()
+    elif mode == 'gt' and cls_num_list is not None:
+        py = np.array(cls_num_list) / cls_num
+    else:
+        raise ValueError("'mode' only support options: 'soft', 'hard', 'gt'")
+
+    return py / py.sum()
+
+
+
+def get_marginal(probs: np.ndarray, cls_num: int, mode: str = 'soft'):
+    r"""
+    Get Marginal Distribution $P(Y=.)$ given $P(Y=.|X=x)$ by summing over x
+    """
+    assert (mode in ['soft', 'hard']) and probs.shape[-1] == cls_num
+    if mode == 'hard':
+        qz = np.zeros(cls_num)
+        for i in np.argmax(probs, axis=-1):
+            qz[i] += 1.
+        qz = qz / qz.sum()
+    elif mode == 'soft':
+        qz = np.mean(probs, axis=0)
+
+    return qz
+
+
+def get_confusion_matrix(probs: np.ndarray,
+                         labels: List,
+                         cls_num: int,
+                         mode: str = 'soft'):
+    r"""
+    Get Confusion Matrix of prediction given prediction $P(Y=.|X=x)$ and ground truth label $Y=.$
+    """
+    assert (mode in ['soft', 'hard']) and probs.shape[-1] == cls_num
+    cm = np.zeros((cls_num, cls_num))
+    if mode == 'soft':
+        for i, j in zip(labels, probs):
+            cm[i, :] += j
+    elif mode == 'hard':
+        labels_pred = np.argmax(probs, axis=-1)
+        for i, j in zip(labels, labels_pred):
+            cm[i, j] += 1
+
+    return cm
+
+
+def normalized(a, axis=-1, order=2):
+    r"""
+    Prediction Normalization
+    """
+    l2 = np.atleast_1d(np.linalg.norm(a, order, axis))
+    l2[l2 == 0] = 1
+    return a / np.expand_dims(l2, axis)
+
+
+def Topk_qy(probs: np.ndarray, cls_num, topk_ratio=0.8, head=0, normalize=True):
+    r"""
+    Get Marginal Distribution $P(Y=.)$ given Topk of $P(Y=.|X=x)$ by summing over x
+    """
+    assert probs.shape[-1] == cls_num
+
+    k = np.clip(int(cls_num * topk_ratio) + head, head + 1, cls_num)
+    qy = np.zeros(cls_num)
+    for x in probs:
+        idx = np.argsort(x)[::-1]
+        idx = idx[head:k]
+        qy[idx] += x[idx]
+
+    if normalize:
+        qy = qy / probs.shape[0]
+
+    return qy

OnZeta/MAPLS/common_cuda.py

@@ -0,0 +1,80 @@
+import torch
+import torch.nn.functional as F
+
+def lsc_torch(probs: torch.Tensor, w: torch.Tensor):
+    """
+    GPU-compatible LSC implementation
+    probs: Tensor [N, C]
+    w: Tensor [C]
+    """
+    assert probs.shape[-1] == w.shape[0], "Shape mismatch"
+
+    weighted_probs = probs * w
+    pc_probs = F.normalize(weighted_probs, p=1, dim=-1)
+    return pc_probs
+def get_py_torch(probs: torch.Tensor, cls_num_list=None, mode='soft'):
+    """
+    GPU-compatible estimation of P(Y_s=i)
+    probs: Tensor [N, C]
+    """
+    cls_num = probs.shape[-1]
+
+    if mode == "soft":
+        py = torch.mean(probs, dim=0)
+    elif mode == "hard":
+        preds = torch.argmax(probs, dim=-1)
+        py = torch.bincount(preds, minlength=cls_num).float()
+        py = py / py.sum()
+    elif mode == 'gt' and cls_num_list is not None:
+        py = torch.tensor(cls_num_list, dtype=torch.float32, device=probs.device)
+        py = py / py.sum()
+    else:
+        raise ValueError("mode must be 'soft', 'hard', or 'gt'")
+
+    return py
+def get_marginal_torch(probs: torch.Tensor, cls_num: int, mode='soft'):
+    assert probs.shape[-1] == cls_num
+    if mode == 'hard':
+        pred = torch.argmax(probs, dim=-1)
+        qz = torch.bincount(pred, minlength=cls_num).float()
+        qz = qz / qz.sum()
+    elif mode == 'soft':
+        qz = torch.mean(probs, dim=0)
+    return qz
+def get_confusion_matrix_torch(probs: torch.Tensor, labels: torch.Tensor, cls_num: int, mode='soft'):
+    """
+    probs: Tensor [N, C]
+    labels: Tensor [N] (long)
+    returns: [cls_num, cls_num] confusion matrix
+    """
+    cm = torch.zeros((cls_num, cls_num), device=probs.device)
+    if mode == 'soft':
+        for i in range(len(labels)):
+            cm[labels[i]] += probs[i]
+    elif mode == 'hard':
+        pred = torch.argmax(probs, dim=-1)
+        for i in range(len(labels)):
+            cm[labels[i], pred[i]] += 1
+    return cm
+def normalized_torch(a: torch.Tensor, axis=-1, order=2):
+    norm = torch.norm(a, p=order, dim=axis, keepdim=True)
+    norm[norm == 0] = 1.0
+    return a / norm
+def topk_qy_torch(probs: torch.Tensor, cls_num: int, topk_ratio=0.8, head=0, normalize=True):
+    """
+    probs: Tensor [N, C]
+    return: Tensor [C]
+    """
+    N, C = probs.shape
+    k = max(min(int(cls_num * topk_ratio) + head, cls_num), head + 1)
+
+    qy = torch.zeros(cls_num, device=probs.device)
+    topk_vals, topk_indices = torch.topk(probs, k=k, dim=1)
+
+    for i in range(N):
+        qy[topk_indices[i][head:]] += topk_vals[i][head:]
+
+    if normalize:
+        qy = qy / N
+
+    return qy

OnZeta/MAPLS/mapls.py

@@ -0,0 +1,153 @@
+import numpy as np
+from .common import normalized, Topk_qy
+import logging
+
+
+def mapls(test_probs,
+          pz: np.ndarray,
+          qy_mode: str = 'soft',
+          max_iter: int = 100,
+          init_mode: str = 'identical',
+          lam: float = None,
+          dvg_name='kl'):
+    r"""
+    Implementation of Maximum A Posteriori Label Shift,
+    for Unknown target label distribution estimation
+
+    Given source domain P(Y_s=i|X_s=x) = f(x) and P(Y_s=i),
+    estimate targe domain P(Y_t=i) on test set
+
+    """
+    # FIXME
+    if test_probs.is_cuda:
+        test_probs = test_probs.cpu()
+    test_probs = test_probs.detach().numpy()
+
+    # Sanity Check
+    cls_num = len(pz)
+    assert test_probs.shape[-1] == cls_num
+    if type(max_iter) != int or max_iter < 0:
+        raise Exception('max_iter should be a positive integer, not ' + str(max_iter))
+
+    # Setup d(p,q) measure
+    if dvg_name == 'kl':
+        dvg = kl_div
+    elif dvg_name == 'js':
+        dvg = js_div
+    else:
+        raise Exception('Unsupported distribution distance measure, expect kl or js.')
+
+    # Set Prior of Target Label Distribution
+    q_prior = np.ones(cls_num) / cls_num
+    # q_prior = pz.copy()
+
+    # Lambda estimation-------------------------------------------------------#
+    if lam is None:
+        # logging.info('Data shape: %s, %s' % (str(train_probs.shape), str(test_probs.shape)))
+        # logging.info('Divergence type is %s' % (dvg))
+        # lam = get_lamda(test_probs, pz, q_prior, dvg=dvg, max_iter=max_iter)  # FIXME why return none
+        lam = lam
+        # logging.info('Estimated lambda value is %.4f' % lam)
+    else:
+        # print('Assigned lambda is %.4f' % lam)
+        pass
+    logging.info("lam is %.4f" % lam)
+
+    # EM Algorithm Computation
+    qz = mapls_EM(test_probs, pz, lam, q_prior, cls_num,
+                  init_mode=init_mode, max_iter=max_iter, qy_mode=qy_mode)
+
+    return qz
+
+
+def mapls_EM(probs, pz, lam, q_prior, cls_num, init_mode='identical', max_iter=100, qy_mode='soft'):
+    # Normalize Source Label Distribution pz
+    pz = np.array(pz) / np.sum(pz)
+    # Initialize Target Label Distribution qz
+    if init_mode == 'uniform':
+        qz = np.ones(cls_num) / cls_num
+    elif init_mode == 'identical':
+        qz = pz.copy()
+    else:
+        raise ValueError('init_mode should be either "uniform" or "identical"')
+
+    # Initialize w
+    w = (np.array(qz) / np.array(pz))
+    # EM algorithm with MAP estimation----------------------------------------#
+    for i in range(max_iter):
+        # print('w shape ', w.shape)
+
+        # E-Step--------------------------------------------------------------#
+        mapls_probs = normalized(probs * w, axis=-1, order=1)
+
+        # M-Step--------------------------------------------------------------#
+        if qy_mode == 'hard':
+            pred = np.argmax(mapls_probs, axis=-1)
+            qz_new = np.bincount(pred.reshape(-1), minlength=cls_num)
+        elif qy_mode == 'soft':
+            qz_new = np.mean(mapls_probs, axis=0)
+        elif qy_mode == 'topk':
+            qz_new = Topk_qy(mapls_probs, cls_num, topk_ratio=0.9, head=0)
+        else:
+            raise Exception('MAPLS mode should be either "soft" or "hard". ')
+        # print(np.shape(pc_probs), np.shape(pred), np.shape(cls_num_list_t))
+
+        # Update w with MAP estimation of Target Label Distribution qz
+        # qz = (qz_new + alpha) / (N + np.sum(alpha))
+        qz = lam * qz_new + (1 - lam) * q_prior
+        qz /= qz.sum()
+        w = qz / pz
+
+    return qz
+
+
+def get_lamda(test_probs, pz, q_prior, dvg, max_iter=50):
+
+
+    K = len(pz)
+
+    # MLLS estimation of source and target domain label distribution
+    qz_pred = mapls_EM(test_probs, pz, 1, 0, K, max_iter=max_iter)
+
+    TU_div = dvg(qz_pred, q_prior)
+    TS_div = dvg(qz_pred, pz)
+    SU_div = dvg(pz, q_prior)
+    print('weights are, TU_div %.4f, TS_div %.4f, SU_div %.4f' % (TU_div, TS_div, SU_div))
+
+    SU_conf = 1 - lam_forward(SU_div, lam_inv(dpq=0.5, lam=0.2))
+    TU_conf = lam_forward(TU_div, lam_inv(dpq=0.5, lam=SU_conf))
+    TS_conf = lam_forward(TS_div, lam_inv(dpq=0.5, lam=SU_conf))
+    print('weights are, unviform_weight %.4f, differ_weight %.4f, regularize weight %.4f'
+                 % (TU_conf, TS_conf, SU_conf))
+
+    confs = np.array([TU_conf, 1 - TS_conf])
+    w = np.array([0.9, 0.1])
+    lam = np.sum(w * confs)
+
+    print('Estimated lambda is: %.4f', lam)
+
+    return lam
+
+
+def lam_inv(dpq, lam):
+    # clip for small lam
+    lam = 1e-3 if abs(lam - 1) < 1e-3 else lam
+    return (1 / (1 - lam) - 1) / dpq
+
+def lam_forward(dpq, gamma):
+    return gamma * dpq / (1 + gamma * dpq)
+
+
+def kl_div(p, q):
+    # fixme
+    # if p == q:
+    #     return 0.0
+    p = np.asarray(p, dtype=np.float16)
+    q = np.asarray(q + 1e-8, dtype=np.float16)
+
+    return np.sum(np.where(p != 0, p * np.log(p / q), 0))
+
+def js_div(p, q):
+    assert (np.abs(np.sum(p) - 1) < 1e-6) and (np.abs(np.sum(q) - 1) < 1e-6)
+    m = (p + q) / 2
+    return kl_div(p, m) / 2 + kl_div(q, m) / 2

OnZeta/MAPLS/mapls_cuda.py

@@ -0,0 +1,76 @@
+import torch
+import torch.nn.functional as F
+
+def mapls_torch(test_probs: torch.Tensor,
+                pz: torch.Tensor,
+                qy_mode: str = 'soft',
+                max_iter: int = 100,
+                init_mode: str = 'identical',
+                lam: float = None,
+                dvg_name='kl') -> torch.Tensor:
+    """
+    GPU-compatible MAP Label Shift (MAPLS) using PyTorch.
+    """
+    device = test_probs.device
+    pz = torch.tensor(pz, dtype=torch.float32, device='cuda')
+    cls_num = pz.numel()
+    assert test_probs.shape[-1] == cls_num
+
+    if dvg_name == 'kl':
+        dvg = kl_div_torch
+    elif dvg_name == 'js':
+        dvg = js_div_torch
+    else:
+        raise ValueError('Unsupported divergence type')
+
+    # Prior: uniform or given
+    q_prior = torch.ones(cls_num, device=device) / cls_num
+
+    # EM
+    qz = mapls_EM_torch(test_probs, pz, lam, q_prior, cls_num,
+                        init_mode=init_mode, max_iter=max_iter, qy_mode=qy_mode)
+    return qz
+
+
+def mapls_EM_torch(probs, pz, lam, q_prior, cls_num, init_mode='identical', max_iter=100, qy_mode='soft'):
+    pz = pz / pz.sum()
+    if init_mode == 'uniform':
+        qz = torch.ones(cls_num, device=probs.device) / cls_num
+    elif init_mode == 'identical':
+        qz = pz.clone()
+    else:
+        raise ValueError('init_mode must be "uniform" or "identical"')
+
+    w = qz / pz
+
+    for _ in range(max_iter):
+        mapls_probs = normalize_torch(probs * w, dim=-1)
+
+        if qy_mode == 'hard':
+            pred = torch.argmax(mapls_probs, dim=-1)
+            qz_new = torch.bincount(pred, minlength=cls_num).float().to(probs.device)
+        elif qy_mode == 'soft':
+            qz_new = mapls_probs.mean(dim=0)
+        else:
+            raise ValueError('qy_mode must be "soft" or "hard"')
+
+        qz = lam * qz_new + (1 - lam) * q_prior
+        qz = qz / qz.sum()
+        w = qz / pz
+
+    return qz
+
+
+def normalize_torch(x, dim=-1, eps=1e-8):
+    return x / (x.sum(dim=dim, keepdim=True) + eps)
+
+
+def kl_div_torch(p, q, eps=1e-8):
+    p = p.to(torch.float32)
+    q = (q + eps).to(torch.float32)
+    return torch.sum(torch.where(p != 0, p * torch.log(p / q), torch.zeros_like(p)))
+
+
+def js_div_torch(p, q):
+    m = (p + q) / 2
+    return (kl_div_torch(p, m) + kl_div_torch(q, m)) / 2

OnZeta/README.md

@@ -0,0 +1,24 @@
+# OnZeta
+PyTorch Implementation for Our ECCV'24 Paper: "Online Zero-Shot Classification with CLIP"
+
+## Requirements
+* Python 3.9
+* PyTorch 1.12
+* [CLIP](https://github.com/openai/CLIP)
+
+## Usage:
+OnZeta with pre-trained ResNet-50
+```
+python main_online.py -a RN50 --data_path /path/to/imagenet
+```
+
+## Citation
+If you use the package in your research, please cite our paper:
+```
+@inproceedings{qian2024onzeta,
+  author    = {Qi Qian and
+               Juhua Hu},
+  title     = {Online Zero-Shot Classification with CLIP},
+  booktitle = {The 18th European Conference on Computer Vision, {ECCV} 2024},
+  year      = {2024}
+}

OnZeta/clip_cifar10.py

@@ -0,0 +1,110 @@
+import torch
+import clip
+import os
+from torchvision.datasets import MNIST, CIFAR10
+import numpy as np
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+model, preprocess = clip.load('RN50', device)
+
+# from https://github.com/openai/CLIP/blob/main/data/prompts.md
+mnist_classes = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9', ]
+mnist_templates = ['a photo of the number: "{}".', ]
+cifar10_classes = ['airplane',
+                   'automobile',
+                   'bird',
+                   'cat',
+                   'deer',
+                   'dog',
+                   'frog',
+                   'horse',
+                   'ship',
+                   'truck', ]
+cifar10_templates = [
+    'a photo of a {}.',
+    'a blurry photo of a {}.',
+    'a black and white photo of a {}.',
+    'a low contrast photo of a {}.',
+    'a high contrast photo of a {}.',
+    'a bad photo of a {}.',
+    'a good photo of a {}.',
+    'a photo of a small {}.',
+    'a photo of a big {}.',
+    'a photo of the {}.',
+    'a blurry photo of the {}.',
+    'a black and white photo of the {}.',
+    'a low contrast photo of the {}.',
+    'a high contrast photo of the {}.',
+    'a bad photo of the {}.',
+    'a good photo of the {}.',
+    'a photo of the small {}.',
+    'a photo of the big {}.',
+]
+
+class_map = {'MNIST': mnist_classes, 'CIFAR10': cifar10_classes}
+template_map = {'MNIST': mnist_templates, 'CIFAR10': cifar10_templates}
+
+
+@torch.no_grad()
+def accuracy(output, target, topk=(1,)):
+    maxk = max(topk)
+    batch_size = target.size(0)
+
+    _, pred = output.topk(maxk, 1, True, True)
+    pred = pred.t()
+    correct = pred.eq(target.reshape(1, -1).expand_as(pred))
+
+    res = []
+    for k in topk:
+        correct_k = correct[:k].reshape(-1).float().sum(0, keepdim=True)
+        res.append(correct_k.mul_(100.0 / batch_size).item())
+    return res
+
+
+@torch.no_grad()
+def extract_text_features(dataset_name):
+    # code borrowed from: https://github.com/openai/CLIP/blob/fcab8b6eb92af684e7ff0a904464be7b99b49b88/notebooks/Prompt_Engineering_for_ImageNet.ipynb
+    class_names = class_map[dataset_name]
+    templates = template_map[dataset_name]
+    model.to(device)
+    model.eval()
+
+    zeroshot_weights = []
+    for classname in class_names:
+        texts = [template.format(classname) for template in templates]
+        texts = clip.tokenize(texts).to(device)
+        class_embeddings = model.encode_text(texts)
+        class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+        class_embedding = class_embeddings.mean(dim=0)
+        class_embedding /= class_embedding.norm()
+        zeroshot_weights.append(class_embedding)
+    zeroshot_weights = torch.stack(zeroshot_weights, dim=1).to(device)
+    return zeroshot_weights
+
+
+mnist = MNIST(root=os.path.expanduser("~/.cache"), download=True, train=False)
+cifar10 = CIFAR10(root=os.path.expanduser("~/.cache"), download=True, train=False)
+
+for dataset in [mnist, cifar10]:
+    # extract image feature, code borrowed from: https://github.com/openai/CLIP#zero-shot-prediction
+    image_features = []
+    image_labels = []
+    for image, class_id in dataset:
+        image_input = preprocess(image).unsqueeze(0).to(device)
+        with torch.no_grad():
+            image_feature = model.encode_image(image_input)
+        image_feature /= image_feature.norm()
+        image_features.append(image_feature)
+        image_labels.append(class_id)
+    image_features = torch.stack(image_features, dim=1).to(device)
+    image_features = image_features.squeeze()
+
+    # extract text feature
+    dataset_name = 'MNIST' if dataset == mnist else 'CIFAR10'
+    text_features = extract_text_features(dataset_name)
+
+    # compute top-1 accuracy
+    logits = (100. * image_features @ text_features).softmax(dim=-1)
+    image_labels = torch.tensor(image_labels).unsqueeze(dim=1).to(device)
+    top1_acc = accuracy(logits, image_labels, (1,))
+    print(f'top-1 accuracy for {dataset_name} dataset: {top1_acc[0]:.3f}')

OnZeta/code_draft.py

@@ -0,0 +1,3 @@
+my_list = ["a", "b", "c", "d", "e", "f", "g", "h", "i", "j"]
+
+print(len(my_list))

OnZeta/lame/lame.py

@@ -0,0 +1,153 @@
+import torch
+import torch.jit
+import logging
+from typing import List, Dict
+
+import time
+import torch.nn.functional as F
+
+
+__all__ = ["LAME"]
+
+
+
+class AffinityMatrix:
+
+    def __init__(self, **kwargs):
+        pass
+
+    def __call__(X, **kwargs):
+        raise NotImplementedError
+
+    def is_psd(self, mat):
+        eigenvalues = torch.eig(mat)[0][:, 0].sort(descending=True)[0]
+        return eigenvalues, float((mat == mat.t()).all() and (eigenvalues >= 0).all())
+
+    def symmetrize(self, mat):
+        return 1 / 2 * (mat + mat.t())
+
+
+class kNN_affinity(AffinityMatrix):
+    def __init__(self, knn: int, **kwargs):
+        self.knn = knn
+
+    def __call__(self, X):
+        N = X.size(0)
+        dist = torch.norm(X.unsqueeze(0) - X.unsqueeze(1), dim=-1, p=2)  # [N, N]
+        n_neighbors = min(self.knn + 1, N)
+
+        knn_index = dist.topk(n_neighbors, -1, largest=False).indices[:, 1:]  # [N, knn]
+
+        W = torch.zeros(N, N, device=X.device)
+        W.scatter_(dim=-1, index=knn_index, value=1.0)
+
+        return W
+
+# def rbf_affinity_blockwise(X, sigma=1.0, batch_size=1000):
+#     N = X.size(0)
+#     kernel = torch.zeros(N, N, device=X.device)
+#
+#     for i in range(0, N, batch_size):
+#         x_batch = X[i:i+batch_size]  # [B, D]
+#         dist = torch.cdist(x_batch, X, p=2)  # [B, N]
+#         sim = torch.exp(-dist ** 2 / (2 * sigma ** 2))  # [B, N]
+#         kernel[i:i+batch_size] = sim
+#
+#     return kernel
+
+
+class rbf_affinity(AffinityMatrix):
+    def __init__(self, sigma: float, **kwargs):
+        self.sigma = sigma
+        self.k = kwargs['knn']
+
+    def __call__(self, X):
+
+        N = X.size(0)
+        dist = torch.norm(X.unsqueeze(0) - X.unsqueeze(1), dim=-1, p=2)  # [N, N]
+        n_neighbors = min(self.k, N)
+        kth_dist = dist.topk(k=n_neighbors, dim=-1, largest=False).values[:, -1]  # compute k^th distance for each point, [N, knn + 1]
+        sigma = kth_dist.mean()
+        rbf = torch.exp(- dist ** 2 / (2 * sigma ** 2))
+        # mask = torch.eye(X.size(0)).to(X.device)
+        # rbf = rbf * (1 - mask)
+        return rbf
+
+
+class linear_affinity(AffinityMatrix):
+
+    def __call__(self, X: torch.Tensor):
+        """
+        X: [N, d]
+        """
+        return torch.matmul(X, X.t())
+
+
+class LAME:
+    """
+    Our proposed method based on Laplacian Regularization.
+    """
+    def __init__(self):
+        """
+        Args:
+            cfg (CfgNode):
+        """
+        self.knn = 5
+        self.sigma = 1.0
+        # print("self.knn, self.sigma".format(self.knn, self.sigma))
+        self.LAME_AFFINITY = 'rbf'
+        self.affinity = eval(f'{self.LAME_AFFINITY}_affinity')(sigma=self.sigma, knn=self.knn)
+        self.force_symmetry = False
+
+
+    def run_step(self, logits, feats):
+        # print("self.knn = {}, self.sigma = {}".format(self.knn, self.sigma))
+        with torch.no_grad():
+            probas = logits.softmax(1)  # [N, K]# [N, d]
+
+            # --- Get unary and terms and kernel ---
+
+            unary = - torch.log(probas + 1e-10)  # [N, K]
+            feats = F.normalize(feats, p=2, dim=-1)  # [N, d]
+            kernel = self.affinity(feats)  # [N, N]
+            # kernel = rbf_affinity_blockwise(feats, sigma=self.sigma, batch_size=1000)
+            if self.force_symmetry:
+                kernel = 1/2 * (kernel + kernel.t())
+            # --- Perform optim ---
+            Y = laplacian_optimization(unary, kernel)
+
+        return Y
+
+
+def laplacian_optimization(unary, kernel, bound_lambda=1, max_steps=100):
+
+    E_list = []
+    oldE = float('inf')
+    Y = (-unary).softmax(-1)  # [N, K]
+    device = kernel.device
+    # print("device: ", device)
+    Y = Y.to(device)
+    unary = unary.to(device)
+    for i in range(max_steps):
+        kernel = kernel.to(torch.float16)
+        Y = Y.to(torch.float16)
+        # bound_lambda = bound_lambda.to(device)
+        pairwise = bound_lambda * kernel.matmul(Y)  # [N, K]
+        pairwise = pairwise.to(torch.float16)
+        exponent = -unary + pairwise
+        Y = exponent.softmax(-1)
+        E = entropy_energy(Y, unary, pairwise, bound_lambda).item()
+        E_list.append(E)
+
+        if (i > 1 and (abs(E - oldE) <= 1e-8 * abs(oldE))):
+            break
+        else:
+            oldE = E
+
+    return Y
+
+
+def entropy_energy(Y, unary, pairwise, bound_lambda):
+    # torch.cuda.empty_cache()
+    E = (unary * Y - bound_lambda * pairwise * Y + Y * torch.log(Y.clip(1e-20))).sum()
+    return E

OnZeta/lame/lame_px.py

@@ -0,0 +1,71 @@
+import torch
+import torch.nn.functional as F
+
+class AffinityMatrix:
+
+    def __init__(self, **kwargs):
+        pass
+
+    def __call__(X, **kwargs):
+        raise NotImplementedError
+
+    def is_psd(self, mat):
+        eigenvalues = torch.eig(mat)[0][:, 0].sort(descending=True)[0]
+        return eigenvalues, float((mat == mat.t()).all() and (eigenvalues >= 0).all())
+
+    def symmetrize(self, mat):
+        return 1 / 2 * (mat + mat.t())
+
+class kNN_affinity(AffinityMatrix):
+    def __init__(self, knn: int, **kwargs):
+        self.knn = knn
+
+    def __call__(self, X):
+        N = X.size(0)
+        dist = torch.norm(X.unsqueeze(0) - X.unsqueeze(1), dim=-1, p=2)  # [N, N]
+        n_neighbors = min(self.knn + 1, N)
+
+        knn_index = dist.topk(n_neighbors, -1, largest=False).indices[:, 1:]  # [N, knn]
+
+        W = torch.zeros(N, N, device=X.device)
+        W.scatter_(dim=-1, index=knn_index, value=1.0)
+
+        return W
+
+def laplacian_optimization(unary, kernel, bound_lambda=1, max_steps=100):
+
+    E_list = []
+    oldE = float('inf')
+    Y = (-unary).softmax(-1)  # [N, K]
+    for i in range(max_steps):
+        pairwise = bound_lambda * kernel.matmul(Y)  # [N, K]
+        exponent = -unary + pairwise
+        Y = exponent.softmax(-1)
+        E = entropy_energy(Y, unary, pairwise, bound_lambda).item()
+        E_list.append(E)
+
+        if (i > 1 and (abs(E - oldE) <= 1e-8 * abs(oldE))):
+            # logger.info(f'Converged in {i} iterations')
+            break
+        else:
+            oldE = E
+
+    return Y
+
+def entropy_energy(Y, unary, pairwise, bound_lambda):
+    E = (unary * Y - bound_lambda * pairwise * Y + Y * torch.log(Y.clip(1e-20))).sum()
+    return E
+
+def lame(probas, feats):
+    unary = - torch.log(probas + 1e-10)  # [N, K]
+
+    # feats = self.model.backbone.out  # [N, d]
+    feats = F.normalize(feats, p=2, dim=-1)  # [N, d]
+    affinity = kNN_affinity(1)
+    kernel = affinity(feats)  # [N, N]
+    if False:
+        kernel = 1 / 2 * (kernel + kernel.t())
+    Y = laplacian_optimization(unary, kernel)
+
+    # final_output = format_result(batched_inputs, Y)
+    return Y

OnZeta/logs/debug_onzeta_eval_2025-06-06_22-11-27.log

@@ -0,0 +1,3 @@
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data

OnZeta/logs/debug_onzeta_eval_2025-06-07_00-13-37.log

@@ -0,0 +1,140 @@
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.03
+mean acc of MAPLS is: 91.81
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.9000
+lam is 0.9000
+lam is 0.9000
+lam is 0.9000
+lam is 0.9000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.94
+mean acc of MAPLS is: 91.68
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.01
+mean acc of MAPLS is: 91.57
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.98
+mean acc of MAPLS is: 91.48
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.6000
+lam is 0.6000
+lam is 0.6000
+lam is 0.6000
+lam is 0.6000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.02
+mean acc of MAPLS is: 91.41
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.5000
+lam is 0.5000
+lam is 0.5000
+lam is 0.5000
+lam is 0.5000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.01
+mean acc of MAPLS is: 91.36
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.4000
+lam is 0.4000
+lam is 0.4000
+lam is 0.4000
+lam is 0.4000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.99
+mean acc of MAPLS is: 91.26
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.3000
+lam is 0.3000
+lam is 0.3000
+lam is 0.3000
+lam is 0.3000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.00
+mean acc of MAPLS is: 91.19
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.2000
+lam is 0.2000
+lam is 0.2000
+lam is 0.2000
+lam is 0.2000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.95
+mean acc of MAPLS is: 91.08
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 0.1000
+lam is 0.1000
+lam is 0.1000
+lam is 0.1000
+lam is 0.1000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.98
+mean acc of MAPLS is: 91.05

OnZeta/logs/debug_onzeta_eval_2025-06-11_22-30-48.log

@@ -0,0 +1,150 @@
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.22
+mean acc of onzeta is: 47.75
+mean acc of MAPLS is: 47.92
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 46.34
+mean acc of onzeta is: 46.80
+mean acc of MAPLS is: 48.02
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.8
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 45.14
+mean acc of onzeta is: 45.29
+mean acc of MAPLS is: 47.54
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.7
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 44.03
+mean acc of onzeta is: 44.15
+mean acc of MAPLS is: 47.49
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.6
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 43.26
+mean acc of onzeta is: 43.33
+mean acc of MAPLS is: 47.49
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.5
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 42.73
+mean acc of onzeta is: 42.65
+mean acc of MAPLS is: 47.27
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.4
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 42.27
+mean acc of onzeta is: 42.17
+mean acc of MAPLS is: 47.02
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.3
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 42.01
+mean acc of onzeta is: 41.70
+mean acc of MAPLS is: 46.97
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.2
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 41.94
+mean acc of onzeta is: 41.62
+mean acc of MAPLS is: 46.95
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.1
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 41.92
+mean acc of onzeta is: 41.70
+mean acc of MAPLS is: 46.98

OnZeta/logs/debug_onzeta_eval_2025-06-11_22-37-19.log

@@ -0,0 +1,150 @@
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.59
+mean acc of onzeta is: 71.10
+mean acc of MAPLS is: 71.15
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.31
+mean acc of onzeta is: 70.88
+mean acc of MAPLS is: 71.22
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.8
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.00
+mean acc of onzeta is: 70.34
+mean acc of MAPLS is: 71.51
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.7
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 69.62
+mean acc of onzeta is: 69.92
+mean acc of MAPLS is: 71.36
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.6
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 69.18
+mean acc of onzeta is: 69.56
+mean acc of MAPLS is: 71.50
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.5
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 68.79
+mean acc of onzeta is: 69.11
+mean acc of MAPLS is: 71.25
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.4
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 68.61
+mean acc of onzeta is: 68.86
+mean acc of MAPLS is: 71.20
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.3
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 68.38
+mean acc of onzeta is: 68.65
+mean acc of MAPLS is: 71.21
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.2
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 68.29
+mean acc of onzeta is: 68.48
+mean acc of MAPLS is: 71.13
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the alpha is 0.1
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 68.27
+mean acc of onzeta is: 68.41
+mean acc of MAPLS is: 71.19

OnZeta/logs/debug_onzeta_eval_2025-06-11_22-52-28.log

@@ -0,0 +1,150 @@
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.79
+mean acc of onzeta is: 70.42
+mean acc of MAPLS is: 70.47
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.68
+mean acc of onzeta is: 70.86
+mean acc of MAPLS is: 70.96
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.8
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.82
+mean acc of onzeta is: 71.23
+mean acc of MAPLS is: 71.28
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.7
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.60
+mean acc of onzeta is: 71.14
+mean acc of MAPLS is: 71.15
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.6
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.70
+mean acc of onzeta is: 71.37
+mean acc of MAPLS is: 71.38
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.5
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.76
+mean acc of onzeta is: 71.35
+mean acc of MAPLS is: 71.42
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.4
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.58
+mean acc of onzeta is: 71.13
+mean acc of MAPLS is: 71.16
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.3
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.76
+mean acc of onzeta is: 71.27
+mean acc of MAPLS is: 71.29
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.2
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.68
+mean acc of onzeta is: 70.98
+mean acc of MAPLS is: 71.05
+Namespace(data_path='./CIFAR100_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.1
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 68.27
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 70.70
+mean acc of onzeta is: 70.82
+mean acc of MAPLS is: 70.83

OnZeta/logs/debug_onzeta_eval_2025-06-11_23-00-32.log

@@ -0,0 +1,150 @@
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.15
+mean acc of onzeta is: 47.21
+mean acc of MAPLS is: 47.43
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.26
+mean acc of onzeta is: 47.50
+mean acc of MAPLS is: 47.76
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.8
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.10
+mean acc of onzeta is: 47.42
+mean acc of MAPLS is: 47.66
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.7
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.25
+mean acc of onzeta is: 47.74
+mean acc of MAPLS is: 48.04
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.6
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.16
+mean acc of onzeta is: 47.85
+mean acc of MAPLS is: 48.14
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.5
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.35
+mean acc of onzeta is: 47.93
+mean acc of MAPLS is: 48.12
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.4
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.28
+mean acc of onzeta is: 47.90
+mean acc of MAPLS is: 48.06
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.3
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.28
+mean acc of onzeta is: 47.76
+mean acc of MAPLS is: 47.94
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.2
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.22
+mean acc of onzeta is: 47.54
+mean acc of MAPLS is: 47.73
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+the beta is 0.1
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.22
+mean acc of onzeta is: 47.39
+mean acc of MAPLS is: 47.58

OnZeta/logs/mapls_inloop_mapls_only_RN50.log

@@ -0,0 +1,10 @@
+mean acc of MAPLS only in-loop with lambda 0.95 is: 59.89
+mean acc of MAPLS only in-loop with lambda 0.90 is: 61.17
+mean acc of MAPLS only in-loop with lambda 0.80 is: 61.77
+mean acc of MAPLS only in-loop with lambda 0.70 is: 61.77
+mean acc of MAPLS only in-loop with lambda 0.60 is: 61.73
+mean acc of MAPLS only in-loop with lambda 0.50 is: 61.57
+mean acc of MAPLS only in-loop with lambda 0.40 is: 61.36
+mean acc of MAPLS only in-loop with lambda 0.30 is: 61.12
+mean acc of MAPLS only in-loop with lambda 0.20 is: 60.82
+mean acc of MAPLS only in-loop with lambda 0.10 is: 60.57

OnZeta/logs/mapls_inloop_mapls_only_vitb16.log

@@ -0,0 +1,10 @@
+mean acc of MAPLS only in-loop with lambda 0.95 is: 68.66
+mean acc of MAPLS only in-loop with lambda 0.90 is: 69.61
+mean acc of MAPLS only in-loop with lambda 0.80 is: 70.07
+mean acc of MAPLS only in-loop with lambda 0.70 is: 70.13
+mean acc of MAPLS only in-loop with lambda 0.60 is: 70.02
+mean acc of MAPLS only in-loop with lambda 0.50 is: 69.78
+mean acc of MAPLS only in-loop with lambda 0.40 is: 69.59
+mean acc of MAPLS only in-loop with lambda 0.30 is: 69.35
+mean acc of MAPLS only in-loop with lambda 0.20 is: 69.12
+mean acc of MAPLS only in-loop with lambda 0.10 is: 68.95

OnZeta/logs/onzeta_eval.log

@@ -0,0 +1,191 @@
+2025-06-06 13:07:46,802 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:07:46,802 - INFO - load pre-trained model
+2025-06-06 13:07:48,023 - INFO - load data
+2025-06-06 13:07:53,313 - INFO - obtain text proxy
+2025-06-06 13:07:53,743 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:07:53,743 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:08:00,391 - INFO - mean acc of onlab is: 47.16
+2025-06-06 13:08:00,392 - INFO - mean acc of onzeta is: 47.71
+2025-06-06 13:08:00,392 - INFO - mean acc of MAPLS is: 47.22
+2025-06-06 13:08:00,393 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:08:00,393 - INFO - load pre-trained model
+2025-06-06 13:08:01,549 - INFO - load data
+2025-06-06 13:08:06,638 - INFO - obtain text proxy
+2025-06-06 13:08:06,978 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:08:06,978 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:08:13,675 - INFO - mean acc of onlab is: 47.13
+2025-06-06 13:08:13,676 - INFO - mean acc of onzeta is: 47.75
+2025-06-06 13:08:13,676 - INFO - mean acc of MAPLS is: 47.99
+2025-06-06 13:08:13,677 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:08:13,677 - INFO - load pre-trained model
+2025-06-06 13:08:14,840 - INFO - load data
+2025-06-06 13:08:19,957 - INFO - obtain text proxy
+2025-06-06 13:08:20,296 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:08:20,296 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:08:27,187 - INFO - mean acc of onlab is: 47.25
+2025-06-06 13:08:27,187 - INFO - mean acc of onzeta is: 47.80
+2025-06-06 13:08:27,187 - INFO - mean acc of MAPLS is: 48.03
+2025-06-06 13:08:27,188 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:08:27,188 - INFO - load pre-trained model
+2025-06-06 13:08:28,303 - INFO - load data
+2025-06-06 13:08:33,628 - INFO - obtain text proxy
+2025-06-06 13:08:34,055 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:08:34,056 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:08:40,606 - INFO - mean acc of onlab is: 47.19
+2025-06-06 13:08:40,606 - INFO - mean acc of onzeta is: 47.84
+2025-06-06 13:08:40,606 - INFO - mean acc of MAPLS is: 48.05
+2025-06-06 13:08:40,608 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:08:40,608 - INFO - load pre-trained model
+2025-06-06 13:08:41,717 - INFO - load data
+2025-06-06 13:08:46,808 - INFO - obtain text proxy
+2025-06-06 13:08:47,165 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:08:47,165 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:08:53,823 - INFO - mean acc of onlab is: 47.24
+2025-06-06 13:08:53,823 - INFO - mean acc of onzeta is: 47.78
+2025-06-06 13:08:53,823 - INFO - mean acc of MAPLS is: 47.97
+2025-06-06 13:08:53,825 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:08:53,825 - INFO - load pre-trained model
+2025-06-06 13:08:54,935 - INFO - load data
+2025-06-06 13:09:00,059 - INFO - obtain text proxy
+2025-06-06 13:09:00,403 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:09:00,403 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:09:43,969 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:09:43,970 - INFO - load pre-trained model
+2025-06-06 13:09:45,268 - INFO - load data
+2025-06-06 13:09:50,604 - INFO - obtain text proxy
+2025-06-06 13:09:51,035 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:09:51,035 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:09:52,108 - INFO - lam is 1.0000
+2025-06-06 13:09:53,381 - INFO - lam is 1.0000
+2025-06-06 13:09:54,673 - INFO - lam is 1.0000
+2025-06-06 13:09:55,950 - INFO - lam is 1.0000
+2025-06-06 13:09:57,265 - INFO - lam is 1.0000
+2025-06-06 13:09:57,514 - INFO - mean acc of onlab is: 47.30
+2025-06-06 13:09:57,514 - INFO - mean acc of onzeta is: 47.94
+2025-06-06 13:09:57,514 - INFO - mean acc of MAPLS is: 46.68
+2025-06-06 13:09:57,516 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:09:57,516 - INFO - load pre-trained model
+2025-06-06 13:09:58,668 - INFO - load data
+2025-06-06 13:10:03,793 - INFO - obtain text proxy
+2025-06-06 13:10:04,134 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:10:04,134 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:10:05,219 - INFO - lam is 0.9000
+2025-06-06 13:10:06,732 - INFO - lam is 0.9000
+2025-06-06 13:10:08,073 - INFO - lam is 0.9000
+2025-06-06 13:10:09,377 - INFO - lam is 0.9000
+2025-06-06 13:10:10,674 - INFO - lam is 0.9000
+2025-06-06 13:10:10,905 - INFO - mean acc of onlab is: 47.14
+2025-06-06 13:10:10,905 - INFO - mean acc of onzeta is: 47.76
+2025-06-06 13:10:10,905 - INFO - mean acc of MAPLS is: 47.99
+2025-06-06 13:10:10,907 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:10:10,907 - INFO - load pre-trained model
+2025-06-06 13:10:12,048 - INFO - load data
+2025-06-06 13:10:17,166 - INFO - obtain text proxy
+2025-06-06 13:10:17,507 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:10:17,507 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:10:18,564 - INFO - lam is 0.8000
+2025-06-06 13:10:19,846 - INFO - lam is 0.8000
+2025-06-06 13:10:21,148 - INFO - lam is 0.8000
+2025-06-06 13:10:22,449 - INFO - lam is 0.8000
+2025-06-06 13:10:23,749 - INFO - lam is 0.8000
+2025-06-06 13:10:23,988 - INFO - mean acc of onlab is: 47.26
+2025-06-06 13:10:23,988 - INFO - mean acc of onzeta is: 47.89
+2025-06-06 13:10:23,988 - INFO - mean acc of MAPLS is: 48.11
+2025-06-06 13:10:23,989 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:10:23,989 - INFO - load pre-trained model
+2025-06-06 13:10:25,097 - INFO - load data
+2025-06-06 13:10:30,237 - INFO - obtain text proxy
+2025-06-06 13:10:30,577 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:10:30,577 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:10:31,634 - INFO - lam is 0.7000
+2025-06-06 13:10:32,915 - INFO - lam is 0.7000
+2025-06-06 13:10:34,212 - INFO - lam is 0.7000
+2025-06-06 13:10:35,508 - INFO - lam is 0.7000
+2025-06-06 13:10:36,791 - INFO - lam is 0.7000
+2025-06-06 13:10:37,026 - INFO - mean acc of onlab is: 47.29
+2025-06-06 13:10:37,026 - INFO - mean acc of onzeta is: 47.93
+2025-06-06 13:10:37,027 - INFO - mean acc of MAPLS is: 48.16
+2025-06-06 13:10:37,028 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:10:37,028 - INFO - load pre-trained model
+2025-06-06 13:10:38,132 - INFO - load data
+2025-06-06 13:10:43,278 - INFO - obtain text proxy
+2025-06-06 13:10:43,617 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:10:43,617 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:10:44,672 - INFO - lam is 0.6000
+2025-06-06 13:10:45,942 - INFO - lam is 0.6000
+2025-06-06 13:10:47,208 - INFO - lam is 0.6000
+2025-06-06 13:10:48,522 - INFO - lam is 0.6000
+2025-06-06 13:10:49,795 - INFO - lam is 0.6000
+2025-06-06 13:10:50,025 - INFO - mean acc of onlab is: 47.23
+2025-06-06 13:10:50,025 - INFO - mean acc of onzeta is: 47.85
+2025-06-06 13:10:50,025 - INFO - mean acc of MAPLS is: 48.05
+2025-06-06 13:10:50,026 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:10:50,026 - INFO - load pre-trained model
+2025-06-06 13:10:51,131 - INFO - load data
+2025-06-06 13:10:56,399 - INFO - obtain text proxy
+2025-06-06 13:10:56,743 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:10:56,743 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:10:57,813 - INFO - lam is 0.5000
+2025-06-06 13:10:59,158 - INFO - lam is 0.5000
+2025-06-06 13:11:00,524 - INFO - lam is 0.5000
+2025-06-06 13:11:01,805 - INFO - lam is 0.5000
+2025-06-06 13:11:03,098 - INFO - lam is 0.5000
+2025-06-06 13:11:03,323 - INFO - mean acc of onlab is: 47.14
+2025-06-06 13:11:03,323 - INFO - mean acc of onzeta is: 47.66
+2025-06-06 13:11:03,323 - INFO - mean acc of MAPLS is: 47.76
+2025-06-06 13:11:03,324 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:11:03,324 - INFO - load pre-trained model
+2025-06-06 13:11:04,430 - INFO - load data
+2025-06-06 13:11:09,583 - INFO - obtain text proxy
+2025-06-06 13:11:09,928 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:11:09,928 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:11:11,029 - INFO - lam is 0.4000
+2025-06-06 13:11:12,338 - INFO - lam is 0.4000
+2025-06-06 13:11:13,668 - INFO - lam is 0.4000
+2025-06-06 13:11:14,940 - INFO - lam is 0.4000
+2025-06-06 13:11:16,244 - INFO - lam is 0.4000
+2025-06-06 13:11:16,479 - INFO - mean acc of onlab is: 47.24
+2025-06-06 13:11:16,479 - INFO - mean acc of onzeta is: 47.80
+2025-06-06 13:11:16,479 - INFO - mean acc of MAPLS is: 47.93
+2025-06-06 13:11:16,480 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:11:16,480 - INFO - load pre-trained model
+2025-06-06 13:11:17,643 - INFO - load data
+2025-06-06 13:11:22,919 - INFO - obtain text proxy
+2025-06-06 13:11:23,264 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:11:23,264 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:11:24,336 - INFO - lam is 0.3000
+2025-06-06 13:11:25,793 - INFO - lam is 0.3000
+2025-06-06 13:11:27,266 - INFO - lam is 0.3000
+2025-06-06 13:11:28,757 - INFO - lam is 0.3000
+2025-06-06 13:11:30,128 - INFO - lam is 0.3000
+2025-06-06 13:11:30,367 - INFO - mean acc of onlab is: 47.19
+2025-06-06 13:11:30,367 - INFO - mean acc of onzeta is: 47.79
+2025-06-06 13:11:30,367 - INFO - mean acc of MAPLS is: 47.90
+2025-06-06 13:11:30,368 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:11:30,368 - INFO - load pre-trained model
+2025-06-06 13:11:31,495 - INFO - load data
+2025-06-06 13:11:36,668 - INFO - obtain text proxy
+2025-06-06 13:11:37,008 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:11:37,008 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:11:38,074 - INFO - lam is 0.2000
+2025-06-06 13:11:39,405 - INFO - lam is 0.2000
+2025-06-06 13:11:40,805 - INFO - lam is 0.2000
+2025-06-06 13:11:42,294 - INFO - lam is 0.2000
+2025-06-06 13:11:43,671 - INFO - lam is 0.2000
+2025-06-06 13:11:43,910 - INFO - mean acc of onlab is: 47.40
+2025-06-06 13:11:43,910 - INFO - mean acc of onzeta is: 47.88
+2025-06-06 13:11:43,910 - INFO - mean acc of MAPLS is: 47.95
+2025-06-06 13:11:43,912 - INFO - Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.4, repeat=5)
+2025-06-06 13:11:43,912 - INFO - load pre-trained model
+2025-06-06 13:11:45,020 - INFO - load data
+2025-06-06 13:11:50,158 - INFO - obtain text proxy
+2025-06-06 13:11:50,503 - INFO - accuracy with text proxy: 41.91
+2025-06-06 13:11:50,503 - INFO - online zero-shot transfer: repeat 5 times
+2025-06-06 13:11:51,571 - INFO - lam is 0.1000
+2025-06-06 13:11:52,891 - INFO - lam is 0.1000
+2025-06-06 13:11:54,162 - INFO - lam is 0.1000
+2025-06-06 13:11:55,477 - INFO - lam is 0.1000
+2025-06-06 13:11:56,752 - INFO - lam is 0.1000
+2025-06-06 13:11:57,004 - INFO - mean acc of onlab is: 47.21
+2025-06-06 13:11:57,004 - INFO - mean acc of onzeta is: 47.72
+2025-06-06 13:11:57,004 - INFO - mean acc of MAPLS is: 47.76

OnZeta/logs/onzeta_eval_2025-06-06_13-25-22.log

@@ -0,0 +1,140 @@
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 47.24
+mean acc of onzeta is: 47.65
+mean acc of MAPLS is: 46.42
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.9000
+lam is 0.9000
+lam is 0.9000
+lam is 0.9000
+lam is 0.9000
+mean acc of onlab is: 47.28
+mean acc of onzeta is: 47.53
+mean acc of MAPLS is: 47.94
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 47.22
+mean acc of onzeta is: 47.69
+mean acc of MAPLS is: 47.97
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+lam is 0.7000
+mean acc of onlab is: 47.31
+mean acc of onzeta is: 47.68
+mean acc of MAPLS is: 47.92
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.6000
+lam is 0.6000
+lam is 0.6000
+lam is 0.6000
+lam is 0.6000
+mean acc of onlab is: 47.22
+mean acc of onzeta is: 47.51
+mean acc of MAPLS is: 47.74
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.5000
+lam is 0.5000
+lam is 0.5000
+lam is 0.5000
+lam is 0.5000
+mean acc of onlab is: 47.23
+mean acc of onzeta is: 47.80
+mean acc of MAPLS is: 47.93
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.4000
+lam is 0.4000
+lam is 0.4000
+lam is 0.4000
+lam is 0.4000
+mean acc of onlab is: 47.20
+mean acc of onzeta is: 47.51
+mean acc of MAPLS is: 47.63
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.3000
+lam is 0.3000
+lam is 0.3000
+lam is 0.3000
+lam is 0.3000
+mean acc of onlab is: 47.19
+mean acc of onzeta is: 47.48
+mean acc of MAPLS is: 47.55
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.2000
+lam is 0.2000
+lam is 0.2000
+lam is 0.2000
+lam is 0.2000
+mean acc of onlab is: 47.29
+mean acc of onzeta is: 47.55
+mean acc of MAPLS is: 47.61
+Namespace(data_path='./CIFAR100_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=1, beta=0.8, repeat=5)
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 41.91
+online zero-shot transfer: repeat 5 times
+lam is 0.1000
+lam is 0.1000
+lam is 0.1000
+lam is 0.1000
+lam is 0.1000
+mean acc of onlab is: 47.07
+mean acc of onzeta is: 47.52
+mean acc of MAPLS is: 47.55

OnZeta/logs/onzeta_eval_2025-06-06_21-54-46.log

@@ -0,0 +1,155 @@
+/home/han321/anaconda3/envs/OnZeta/bin/python /home/han321/projects/OnZeta/main_online_cifar10.py
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+/home/han321/projects/OnZeta/main_online_cifar10.py:184: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. (Deprecated NumPy 1.25.)
+  return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+Assigned lambda is 1.0000
+Assigned lambda is 1.0000
+Assigned lambda is 1.0000
+Assigned lambda is 1.0000
+Assigned lambda is 1.0000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.58
+mean acc of MAPLS is: 10.00
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.9000
+Assigned lambda is 0.9000
+Assigned lambda is 0.9000
+Assigned lambda is 0.9000
+Assigned lambda is 0.9000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.62
+mean acc of MAPLS is: 76.45
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.8000
+Assigned lambda is 0.8000
+Assigned lambda is 0.8000
+Assigned lambda is 0.8000
+Assigned lambda is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.66
+mean acc of MAPLS is: 77.57
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.7000
+Assigned lambda is 0.7000
+Assigned lambda is 0.7000
+Assigned lambda is 0.7000
+Assigned lambda is 0.7000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.63
+mean acc of MAPLS is: 77.22
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.6000
+Assigned lambda is 0.6000
+Assigned lambda is 0.6000
+Assigned lambda is 0.6000
+Assigned lambda is 0.6000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.65
+mean acc of MAPLS is: 76.49
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.5000
+Assigned lambda is 0.5000
+Assigned lambda is 0.5000
+Assigned lambda is 0.5000
+Assigned lambda is 0.5000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.62
+mean acc of MAPLS is: 75.84
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.4000
+Assigned lambda is 0.4000
+Assigned lambda is 0.4000
+Assigned lambda is 0.4000
+Assigned lambda is 0.4000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.58
+mean acc of MAPLS is: 75.02
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.3000
+Assigned lambda is 0.3000
+Assigned lambda is 0.3000
+Assigned lambda is 0.3000
+Assigned lambda is 0.3000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.60
+mean acc of MAPLS is: 74.05
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.2000
+Assigned lambda is 0.2000
+Assigned lambda is 0.2000
+Assigned lambda is 0.2000
+Assigned lambda is 0.2000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.59
+mean acc of MAPLS is: 73.21
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+load pre-trained model
+load data
+Files already downloaded and verified
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+Assigned lambda is 0.1000
+Assigned lambda is 0.1000
+Assigned lambda is 0.1000
+Assigned lambda is 0.1000
+Assigned lambda is 0.1000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.64
+mean acc of MAPLS is: 72.43
+
+Process finished with exit code 0

OnZeta/logs/onzeta_eval_2025-06-11_20-29-37.log

@@ -0,0 +1,150 @@
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.46
+mean acc of onzeta is: 90.71
+mean acc of MAPLS is: 90.72
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.74
+mean acc of onzeta is: 91.00
+mean acc of MAPLS is: 91.11
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.8
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.90
+mean acc of onzeta is: 91.11
+mean acc of MAPLS is: 91.51
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.7
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.88
+mean acc of onzeta is: 91.08
+mean acc of MAPLS is: 91.56
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.6
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.99
+mean acc of onzeta is: 91.16
+mean acc of MAPLS is: 91.74
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.5
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.86
+mean acc of onzeta is: 91.06
+mean acc of MAPLS is: 91.70
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.4
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.94
+mean acc of onzeta is: 91.10
+mean acc of MAPLS is: 91.74
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.3
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.89
+mean acc of onzeta is: 91.12
+mean acc of MAPLS is: 91.72
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.2
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.83
+mean acc of onzeta is: 91.01
+mean acc of MAPLS is: 91.80
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.1
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.78
+mean acc of onzeta is: 90.94
+mean acc of MAPLS is: 91.76

OnZeta/logs/onzeta_eval_2025-06-11_21-19-15.log

@@ -0,0 +1,150 @@
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 75.26
+mean acc of onzeta is: 75.39
+mean acc of MAPLS is: 75.42
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 75.82
+mean acc of onzeta is: 75.90
+mean acc of MAPLS is: 76.36
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.8
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 75.14
+mean acc of onzeta is: 75.17
+mean acc of MAPLS is: 76.65
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.7
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 74.39
+mean acc of onzeta is: 74.41
+mean acc of MAPLS is: 77.04
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.6
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 73.60
+mean acc of onzeta is: 73.69
+mean acc of MAPLS is: 77.21
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.5
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 72.79
+mean acc of onzeta is: 72.76
+mean acc of MAPLS is: 77.44
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.4
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 72.33
+mean acc of onzeta is: 72.37
+mean acc of MAPLS is: 77.46
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.3
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 72.00
+mean acc of onzeta is: 72.02
+mean acc of MAPLS is: 77.54
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.2
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.77
+mean acc of onzeta is: 71.79
+mean acc of MAPLS is: 77.56
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the alpha is 0.1
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.64
+mean acc of onzeta is: 71.64
+mean acc of MAPLS is: 77.60

OnZeta/logs/onzeta_eval_2025-06-11_21-44-32.log

@@ -0,0 +1,150 @@
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.09
+mean acc of MAPLS is: 76.99
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.31
+mean acc of MAPLS is: 77.12
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.8
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.29
+mean acc of MAPLS is: 77.15
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.7
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.50
+mean acc of MAPLS is: 77.31
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.6
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.46
+mean acc of MAPLS is: 77.29
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.5
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.59
+mean acc of MAPLS is: 77.42
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.4
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.60
+mean acc of MAPLS is: 77.55
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.3
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.59
+mean acc of MAPLS is: 77.58
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.2
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.69
+mean acc of MAPLS is: 77.52
+Namespace(data_path='./CIFAR10_TEST', arch='RN50', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.1
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 71.58
+online zero-shot transfer: repeat 5 times
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+lam is 0.8000
+mean acc of onlab is: 71.58
+mean acc of onzeta is: 71.65
+mean acc of MAPLS is: 77.58

OnZeta/logs/onzeta_eval_2025-06-11_22-09-19.log

@@ -0,0 +1,150 @@
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 1.0
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.09
+mean acc of MAPLS is: 91.78
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.9
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.09
+mean acc of MAPLS is: 91.83
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.8
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.03
+mean acc of MAPLS is: 91.82
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.7
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.00
+mean acc of MAPLS is: 91.77
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.6
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 91.04
+mean acc of MAPLS is: 91.78
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.5
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.98
+mean acc of MAPLS is: 91.74
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.4
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.96
+mean acc of MAPLS is: 91.78
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.3
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.94
+mean acc of MAPLS is: 91.75
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.2
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.89
+mean acc of MAPLS is: 91.68
+Namespace(data_path='./CIFAR10_TEST', arch='ViT-B/16', workers=8, batch_size=256, tau_t=0.01, tau_i=0.04, cw=0.5, cr=20, alpha=0, beta=0.4, repeat=5)
+the beta is 0.1
+load pre-trained model
+load data
+obtain text proxy
+accuracy with text proxy: 90.77
+online zero-shot transfer: repeat 5 times
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+lam is 1.0000
+mean acc of onlab is: 90.77
+mean acc of onzeta is: 90.89
+mean acc of MAPLS is: 91.62

OnZeta/main_online_cifar10.py

@@ -0,0 +1,209 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+from torchvision.datasets import MNIST, CIFAR10
+from datetime import datetime
+import logging
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+
+log_filename = os.path.join("logs", f"onzeta_eval_{datetime.now().strftime('%Y-%m-%d_%H-%M-%S')}.log")
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(message)s',
+    handlers=[
+        logging.FileHandler(log_filename),
+        logging.StreamHandler()
+    ]
+)
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='./CIFAR10_TEST', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='ViT-B/16',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=256, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=0, type=float)
+parser.add_argument('--beta', default=0.4, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+device = "cuda" if torch.cuda.is_available() else "cpu"
+
+def main(beta):
+
+    args = parser.parse_args()
+    logging.info(args)
+
+    lam = 1
+    args.beta = beta
+    logging.info("the beta is {}".format(beta))
+
+    cifar10_classes = [
+        'airplane',
+        'automobile',
+        'bird',
+        'cat',
+        'deer',
+        'dog',
+        'frog',
+        'horse',
+        'ship',
+        'truck',
+    ]
+
+    cifar10_templates = [
+        'a photo of a {}.',
+        'a blurry photo of a {}.',
+        'a black and white photo of a {}.',
+        'a low contrast photo of a {}.',
+        'a high contrast photo of a {}.',
+        'a bad photo of a {}.',
+        'a good photo of a {}.',
+        'a photo of a small {}.',
+        'a photo of a big {}.',
+        'a photo of the {}.',
+        'a blurry photo of the {}.',
+        'a black and white photo of the {}.',
+        'a low contrast photo of the {}.',
+        'a high contrast photo of the {}.',
+        'a bad photo of the {}.',
+        'a good photo of the {}.',
+        'a photo of the small {}.',
+        'a photo of the big {}.',
+    ]
+
+
+    logging.info('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    logging.info('load data')
+    # valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    cifar10 = CIFAR10(root=os.path.expanduser("~/.cache"), download=True, train=False)
+    # val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    # loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(cifar10):
+            # images = images.cuda()
+            # target = target.cuda()
+            # image_features = model.encode_image(images)
+            images = preprocess(images).unsqueeze(0).to(device)
+            with torch.no_grad():
+                images = model.encode_image(images)
+            images /= images.norm()
+            image_feat.append(images)
+            image_label.append(target)
+    image_feat = torch.stack(image_feat, dim=1).to(device)
+    image_feat = image_feat.squeeze()
+    # image_label = torch.cat(image_label, dim=0)
+    image_label = torch.tensor(image_label, dtype=torch.long).to(device)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    logging.info('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, cifar10_classes, cifar10_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    logging.info(f'accuracy with text proxy: {top1:.2f}')
+
+    logging.info('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+            beta = args.beta * math.sqrt((i + 1) / n)
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+            rho -= rlr * (tlabel - args.alpha / num_class)
+            rho[rho < 0] = 0
+            text_label[i, :] = tlabel
+            vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            grad = torch.outer(x, vision_label - tlabel)
+            w -= (lr / args.tau_i) * grad
+            w = F.normalize(w, dim=0)
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+
+        # MAPLS - EM Algorithm
+        pz = np.full(len(cifar10_classes), 1.0 / len(cifar10_classes))
+        qy = mapls(combo_label, pz=pz, qy_mode="soft", max_iter=100, lam=lam)  # FIXME why return nan
+
+        w = np.array(qy) / np.array(pz)
+        if combo_label.is_cuda:
+            combo_label_cpu = combo_label.cpu()
+        qy_probs = lsc(combo_label_cpu, 1.0 / w)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+    logging.info('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    logging.info('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    logging.info('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    betas = [1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1]
+    for beta in betas:
+        main(beta)
+

OnZeta/main_online_cifar100.py

@@ -0,0 +1,292 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+from datetime import datetime
+import logging
+
+log_filename = os.path.join("logs", f"debug_onzeta_eval_{datetime.now().strftime('%Y-%m-%d_%H-%M-%S')}.log")
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(message)s',
+    handlers=[
+        logging.FileHandler(log_filename),
+        logging.StreamHandler()
+    ]
+)
+
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='./CIFAR100_TEST', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='RN50',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=256, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.4, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+
+
+def main(beta):
+
+    args = parser.parse_args()
+    logging.info(args)
+
+    args.beta = beta
+    lam = 0.7
+    logging.info("the beta is {}".format(beta))
+
+    cifar100_classes = [
+        'apple',
+        'aquarium fish',
+        'baby',
+        'bear',
+        'beaver',
+        'bed',
+        'bee',
+        'beetle',
+        'bicycle',
+        'bottle',
+        'bowl',
+        'boy',
+        'bridge',
+        'bus',
+        'butterfly',
+        'camel',
+        'can',
+        'castle',
+        'caterpillar',
+        'cattle',
+        'chair',
+        'chimpanzee',
+        'clock',
+        'cloud',
+        'cockroach',
+        'couch',
+        'crab',
+        'crocodile',
+        'cup',
+        'dinosaur',
+        'dolphin',
+        'elephant',
+        'flatfish',
+        'forest',
+        'fox',
+        'girl',
+        'hamster',
+        'house',
+        'kangaroo',
+        'keyboard',
+        'lamp',
+        'lawn mower',
+        'leopard',
+        'lion',
+        'lizard',
+        'lobster',
+        'man',
+        'maple tree',
+        'motorcycle',
+        'mountain',
+        'mouse',
+        'mushroom',
+        'oak tree',
+        'orange',
+        'orchid',
+        'otter',
+        'palm tree',
+        'pear',
+        'pickup truck',
+        'pine tree',
+        'plain',
+        'plate',
+        'poppy',
+        'porcupine',
+        'possum',
+        'rabbit',
+        'raccoon',
+        'ray',
+        'road',
+        'rocket',
+        'rose',
+        'sea',
+        'seal',
+        'shark',
+        'shrew',
+        'skunk',
+        'skyscraper',
+        'snail',
+        'snake',
+        'spider',
+        'squirrel',
+        'streetcar',
+        'sunflower',
+        'sweet pepper',
+        'table',
+        'tank',
+        'telephone',
+        'television',
+        'tiger',
+        'tractor',
+        'train',
+        'trout',
+        'tulip',
+        'turtle',
+        'wardrobe',
+        'whale',
+        'willow tree',
+        'wolf',
+        'woman',
+        'worm',
+    ]
+
+    cifar100_templates = [
+        'a photo of a {}.',
+        'a blurry photo of a {}.',
+        'a black and white photo of a {}.',
+        'a low contrast photo of a {}.',
+        'a high contrast photo of a {}.',
+        'a bad photo of a {}.',
+        'a good photo of a {}.',
+        'a photo of a small {}.',
+        'a photo of a big {}.',
+        'a photo of the {}.',
+        'a blurry photo of the {}.',
+        'a black and white photo of the {}.',
+        'a low contrast photo of the {}.',
+        'a high contrast photo of the {}.',
+        'a bad photo of the {}.',
+        'a good photo of the {}.',
+        'a photo of the small {}.',
+        'a photo of the big {}.',
+    ]
+
+
+    logging.info('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    logging.info('load data')
+    # valdir = os.path.join(args.data_path, 'val')
+    valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    logging.info('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, cifar100_classes, cifar100_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    logging.info(f'accuracy with text proxy: {top1:.2f}')
+
+    logging.info('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+            beta = args.beta * math.sqrt((i + 1) / n)
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+            rho -= rlr * (tlabel - args.alpha / num_class)
+            rho[rho < 0] = 0
+            text_label[i, :] = tlabel
+            vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            grad = torch.outer(x, vision_label - tlabel)
+            w -= (lr / args.tau_i) * grad
+            w = F.normalize(w, dim=0)
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+
+        # MAPLS - EM Algorithm
+        pz = np.full(len(cifar100_classes), 1.0 / len(cifar100_classes))
+        qy = mapls(combo_label, pz=pz, qy_mode="soft", max_iter=100, lam=lam)  # FIXME why return nan
+
+        w = np.array(qy) / np.array(pz)
+        if combo_label.is_cuda:
+            combo_label_cpu = combo_label.cpu()
+        qy_probs = lsc(combo_label_cpu, 1.0 / w)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+    logging.info('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    logging.info('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    logging.info('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    betas = [1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1]
+    for beta in betas:
+        main(beta)
+

OnZeta/main_online_imagenet_adap_freq.py

@@ -0,0 +1,353 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='/home/lt_test/projects/datasets/ImageNet/', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='ViT-L/14@336px',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=512, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.8, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+parser.add_argument('--adpt_weight', default=0.05, type=float)
+
+
+def main(adpt_weight):
+
+    args = parser.parse_args()
+    print(args)
+    print("adpt_weight: ", adpt_weight)
+    imagenet_classes = ["tench", "goldfish", "great white shark", "tiger shark", "hammerhead shark", "electric ray",
+                        "stingray", "rooster", "hen", "ostrich", "brambling", "goldfinch", "house finch", "junco",
+                        "indigo bunting", "American robin", "bulbul", "jay", "magpie", "chickadee", "American dipper",
+                        "kite (bird of prey)", "bald eagle", "vulture", "great grey owl", "fire salamander",
+                        "smooth newt", "newt", "spotted salamander", "axolotl", "American bullfrog", "tree frog",
+                        "tailed frog", "loggerhead sea turtle", "leatherback sea turtle", "mud turtle", "terrapin",
+                        "box turtle", "banded gecko", "green iguana", "Carolina anole",
+                        "desert grassland whiptail lizard", "agama", "frilled-necked lizard", "alligator lizard",
+                        "Gila monster", "European green lizard", "chameleon", "Komodo dragon", "Nile crocodile",
+                        "American alligator", "triceratops", "worm snake", "ring-necked snake",
+                        "eastern hog-nosed snake", "smooth green snake", "kingsnake", "garter snake", "water snake",
+                        "vine snake", "night snake", "boa constrictor", "African rock python", "Indian cobra",
+                        "green mamba", "sea snake", "Saharan horned viper", "eastern diamondback rattlesnake",
+                        "sidewinder rattlesnake", "trilobite", "harvestman", "scorpion", "yellow garden spider",
+                        "barn spider", "European garden spider", "southern black widow", "tarantula", "wolf spider",
+                        "tick", "centipede", "black grouse", "ptarmigan", "ruffed grouse", "prairie grouse", "peafowl",
+                        "quail", "partridge", "african grey parrot", "macaw", "sulphur-crested cockatoo", "lorikeet",
+                        "coucal", "bee eater", "hornbill", "hummingbird", "jacamar", "toucan", "duck",
+                        "red-breasted merganser", "goose", "black swan", "tusker", "echidna", "platypus", "wallaby",
+                        "koala", "wombat", "jellyfish", "sea anemone", "brain coral", "flatworm", "nematode", "conch",
+                        "snail", "slug", "sea slug", "chiton", "chambered nautilus", "Dungeness crab", "rock crab",
+                        "fiddler crab", "red king crab", "American lobster", "spiny lobster", "crayfish", "hermit crab",
+                        "isopod", "white stork", "black stork", "spoonbill", "flamingo", "little blue heron",
+                        "great egret", "bittern bird", "crane bird", "limpkin", "common gallinule", "American coot",
+                        "bustard", "ruddy turnstone", "dunlin", "common redshank", "dowitcher", "oystercatcher",
+                        "pelican", "king penguin", "albatross", "grey whale", "killer whale", "dugong", "sea lion",
+                        "Chihuahua", "Japanese Chin", "Maltese", "Pekingese", "Shih Tzu", "King Charles Spaniel",
+                        "Papillon", "toy terrier", "Rhodesian Ridgeback", "Afghan Hound", "Basset Hound", "Beagle",
+                        "Bloodhound", "Bluetick Coonhound", "Black and Tan Coonhound", "Treeing Walker Coonhound",
+                        "English foxhound", "Redbone Coonhound", "borzoi", "Irish Wolfhound", "Italian Greyhound",
+                        "Whippet", "Ibizan Hound", "Norwegian Elkhound", "Otterhound", "Saluki", "Scottish Deerhound",
+                        "Weimaraner", "Staffordshire Bull Terrier", "American Staffordshire Terrier",
+                        "Bedlington Terrier", "Border Terrier", "Kerry Blue Terrier", "Irish Terrier",
+                        "Norfolk Terrier", "Norwich Terrier", "Yorkshire Terrier", "Wire Fox Terrier",
+                        "Lakeland Terrier", "Sealyham Terrier", "Airedale Terrier", "Cairn Terrier",
+                        "Australian Terrier", "Dandie Dinmont Terrier", "Boston Terrier", "Miniature Schnauzer",
+                        "Giant Schnauzer", "Standard Schnauzer", "Scottish Terrier", "Tibetan Terrier",
+                        "Australian Silky Terrier", "Soft-coated Wheaten Terrier", "West Highland White Terrier",
+                        "Lhasa Apso", "Flat-Coated Retriever", "Curly-coated Retriever", "Golden Retriever",
+                        "Labrador Retriever", "Chesapeake Bay Retriever", "German Shorthaired Pointer", "Vizsla",
+                        "English Setter", "Irish Setter", "Gordon Setter", "Brittany dog", "Clumber Spaniel",
+                        "English Springer Spaniel", "Welsh Springer Spaniel", "Cocker Spaniel", "Sussex Spaniel",
+                        "Irish Water Spaniel", "Kuvasz", "Schipperke", "Groenendael dog", "Malinois", "Briard",
+                        "Australian Kelpie", "Komondor", "Old English Sheepdog", "Shetland Sheepdog", "collie",
+                        "Border Collie", "Bouvier des Flandres dog", "Rottweiler", "German Shepherd Dog", "Dobermann",
+                        "Miniature Pinscher", "Greater Swiss Mountain Dog", "Bernese Mountain Dog",
+                        "Appenzeller Sennenhund", "Entlebucher Sennenhund", "Boxer", "Bullmastiff", "Tibetan Mastiff",
+                        "French Bulldog", "Great Dane", "St. Bernard", "husky", "Alaskan Malamute", "Siberian Husky",
+                        "Dalmatian", "Affenpinscher", "Basenji", "pug", "Leonberger", "Newfoundland dog",
+                        "Great Pyrenees dog", "Samoyed", "Pomeranian", "Chow Chow", "Keeshond", "brussels griffon",
+                        "Pembroke Welsh Corgi", "Cardigan Welsh Corgi", "Toy Poodle", "Miniature Poodle",
+                        "Standard Poodle", "Mexican hairless dog (xoloitzcuintli)", "grey wolf", "Alaskan tundra wolf",
+                        "red wolf or maned wolf", "coyote", "dingo", "dhole", "African wild dog", "hyena", "red fox",
+                        "kit fox", "Arctic fox", "grey fox", "tabby cat", "tiger cat", "Persian cat", "Siamese cat",
+                        "Egyptian Mau", "cougar", "lynx", "leopard", "snow leopard", "jaguar", "lion", "tiger",
+                        "cheetah", "brown bear", "American black bear", "polar bear", "sloth bear", "mongoose",
+                        "meerkat", "tiger beetle", "ladybug", "ground beetle", "longhorn beetle", "leaf beetle",
+                        "dung beetle", "rhinoceros beetle", "weevil", "fly", "bee", "ant", "grasshopper",
+                        "cricket insect", "stick insect", "cockroach", "praying mantis", "cicada", "leafhopper",
+                        "lacewing", "dragonfly", "damselfly", "red admiral butterfly", "ringlet butterfly",
+                        "monarch butterfly", "small white butterfly", "sulphur butterfly", "gossamer-winged butterfly",
+                        "starfish", "sea urchin", "sea cucumber", "cottontail rabbit", "hare", "Angora rabbit",
+                        "hamster", "porcupine", "fox squirrel", "marmot", "beaver", "guinea pig", "common sorrel horse",
+                        "zebra", "pig", "wild boar", "warthog", "hippopotamus", "ox", "water buffalo", "bison",
+                        "ram (adult male sheep)", "bighorn sheep", "Alpine ibex", "hartebeest", "impala (antelope)",
+                        "gazelle", "arabian camel", "llama", "weasel", "mink", "European polecat",
+                        "black-footed ferret", "otter", "skunk", "badger", "armadillo", "three-toed sloth", "orangutan",
+                        "gorilla", "chimpanzee", "gibbon", "siamang", "guenon", "patas monkey", "baboon", "macaque",
+                        "langur", "black-and-white colobus", "proboscis monkey", "marmoset", "white-headed capuchin",
+                        "howler monkey", "titi monkey", "Geoffroy's spider monkey", "common squirrel monkey",
+                        "ring-tailed lemur", "indri", "Asian elephant", "African bush elephant", "red panda",
+                        "giant panda", "snoek fish", "eel", "silver salmon", "rock beauty fish", "clownfish",
+                        "sturgeon", "gar fish", "lionfish", "pufferfish", "abacus", "abaya", "academic gown",
+                        "accordion", "acoustic guitar", "aircraft carrier", "airliner", "airship", "altar", "ambulance",
+                        "amphibious vehicle", "analog clock", "apiary", "apron", "trash can", "assault rifle",
+                        "backpack", "bakery", "balance beam", "balloon", "ballpoint pen", "Band-Aid", "banjo",
+                        "baluster / handrail", "barbell", "barber chair", "barbershop", "barn", "barometer", "barrel",
+                        "wheelbarrow", "baseball", "basketball", "bassinet", "bassoon", "swimming cap", "bath towel",
+                        "bathtub", "station wagon", "lighthouse", "beaker", "military hat (bearskin or shako)",
+                        "beer bottle", "beer glass", "bell tower", "baby bib", "tandem bicycle", "bikini",
+                        "ring binder", "binoculars", "birdhouse", "boathouse", "bobsleigh", "bolo tie", "poke bonnet",
+                        "bookcase", "bookstore", "bottle cap", "hunting bow", "bow tie", "brass memorial plaque", "bra",
+                        "breakwater", "breastplate", "broom", "bucket", "buckle", "bulletproof vest",
+                        "high-speed train", "butcher shop", "taxicab", "cauldron", "candle", "cannon", "canoe",
+                        "can opener", "cardigan", "car mirror", "carousel", "tool kit", "cardboard box / carton",
+                        "car wheel", "automated teller machine", "cassette", "cassette player", "castle", "catamaran",
+                        "CD player", "cello", "mobile phone", "chain", "chain-link fence", "chain mail", "chainsaw",
+                        "storage chest", "chiffonier", "bell or wind chime", "china cabinet", "Christmas stocking",
+                        "church", "movie theater", "cleaver", "cliff dwelling", "cloak", "clogs", "cocktail shaker",
+                        "coffee mug", "coffeemaker", "spiral or coil", "combination lock", "computer keyboard",
+                        "candy store", "container ship", "convertible", "corkscrew", "cornet", "cowboy boot",
+                        "cowboy hat", "cradle", "construction crane", "crash helmet", "crate", "infant bed",
+                        "Crock Pot", "croquet ball", "crutch", "cuirass", "dam", "desk", "desktop computer",
+                        "rotary dial telephone", "diaper", "digital clock", "digital watch", "dining table",
+                        "dishcloth", "dishwasher", "disc brake", "dock", "dog sled", "dome", "doormat", "drilling rig",
+                        "drum", "drumstick", "dumbbell", "Dutch oven", "electric fan", "electric guitar",
+                        "electric locomotive", "entertainment center", "envelope", "espresso machine", "face powder",
+                        "feather boa", "filing cabinet", "fireboat", "fire truck", "fire screen", "flagpole", "flute",
+                        "folding chair", "football helmet", "forklift", "fountain", "fountain pen", "four-poster bed",
+                        "freight car", "French horn", "frying pan", "fur coat", "garbage truck",
+                        "gas mask or respirator", "gas pump", "goblet", "go-kart", "golf ball", "golf cart", "gondola",
+                        "gong", "gown", "grand piano", "greenhouse", "radiator grille", "grocery store", "guillotine",
+                        "hair clip", "hair spray", "half-track", "hammer", "hamper", "hair dryer", "hand-held computer",
+                        "handkerchief", "hard disk drive", "harmonica", "harp", "combine harvester", "hatchet",
+                        "holster", "home theater", "honeycomb", "hook", "hoop skirt", "gymnastic horizontal bar",
+                        "horse-drawn vehicle", "hourglass", "iPod", "clothes iron", "carved pumpkin", "jeans", "jeep",
+                        "T-shirt", "jigsaw puzzle", "rickshaw", "joystick", "kimono", "knee pad", "knot", "lab coat",
+                        "ladle", "lampshade", "laptop computer", "lawn mower", "lens cap", "letter opener", "library",
+                        "lifeboat", "lighter", "limousine", "ocean liner", "lipstick", "slip-on shoe", "lotion",
+                        "music speaker", "loupe magnifying glass", "sawmill", "magnetic compass", "messenger bag",
+                        "mailbox", "tights", "one-piece bathing suit", "manhole cover", "maraca", "marimba", "mask",
+                        "matchstick", "maypole", "maze", "measuring cup", "medicine cabinet", "megalith", "microphone",
+                        "microwave oven", "military uniform", "milk can", "minibus", "miniskirt", "minivan", "missile",
+                        "mitten", "mixing bowl", "mobile home", "ford model t", "modem", "monastery", "monitor",
+                        "moped", "mortar and pestle", "graduation cap", "mosque", "mosquito net", "vespa",
+                        "mountain bike", "tent", "computer mouse", "mousetrap", "moving van", "muzzle", "metal nail",
+                        "neck brace", "necklace", "baby pacifier", "notebook computer", "obelisk", "oboe", "ocarina",
+                        "odometer", "oil filter", "pipe organ", "oscilloscope", "overskirt", "bullock cart",
+                        "oxygen mask", "product packet / packaging", "paddle", "paddle wheel", "padlock", "paintbrush",
+                        "pajamas", "palace", "pan flute", "paper towel", "parachute", "parallel bars", "park bench",
+                        "parking meter", "railroad car", "patio", "payphone", "pedestal", "pencil case",
+                        "pencil sharpener", "perfume", "Petri dish", "photocopier", "plectrum", "Pickelhaube",
+                        "picket fence", "pickup truck", "pier", "piggy bank", "pill bottle", "pillow", "ping-pong ball",
+                        "pinwheel", "pirate ship", "drink pitcher", "block plane", "planetarium", "plastic bag",
+                        "plate rack", "farm plow", "plunger", "Polaroid camera", "pole", "police van", "poncho",
+                        "pool table", "soda bottle", "plant pot", "potter's wheel", "power drill", "prayer rug",
+                        "printer", "prison", "missile", "projector", "hockey puck", "punching bag", "purse", "quill",
+                        "quilt", "race car", "racket", "radiator", "radio", "radio telescope", "rain barrel",
+                        "recreational vehicle", "fishing casting reel", "reflex camera", "refrigerator",
+                        "remote control", "restaurant", "revolver", "rifle", "rocking chair", "rotisserie", "eraser",
+                        "rugby ball", "ruler measuring stick", "sneaker", "safe", "safety pin", "salt shaker", "sandal",
+                        "sarong", "saxophone", "scabbard", "weighing scale", "school bus", "schooner", "scoreboard",
+                        "CRT monitor", "screw", "screwdriver", "seat belt", "sewing machine", "shield", "shoe store",
+                        "shoji screen / room divider", "shopping basket", "shopping cart", "shovel", "shower cap",
+                        "shower curtain", "ski", "balaclava ski mask", "sleeping bag", "slide rule", "sliding door",
+                        "slot machine", "snorkel", "snowmobile", "snowplow", "soap dispenser", "soccer ball", "sock",
+                        "solar thermal collector", "sombrero", "soup bowl", "keyboard space bar", "space heater",
+                        "space shuttle", "spatula", "motorboat", "spider web", "spindle", "sports car", "spotlight",
+                        "stage", "steam locomotive", "through arch bridge", "steel drum", "stethoscope", "scarf",
+                        "stone wall", "stopwatch", "stove", "strainer", "tram", "stretcher", "couch", "stupa",
+                        "submarine", "suit", "sundial", "sunglasses", "sunglasses", "sunscreen", "suspension bridge",
+                        "mop", "sweatshirt", "swim trunks / shorts", "swing", "electrical switch", "syringe",
+                        "table lamp", "tank", "tape player", "teapot", "teddy bear", "television", "tennis ball",
+                        "thatched roof", "front curtain", "thimble", "threshing machine", "throne", "tile roof",
+                        "toaster", "tobacco shop", "toilet seat", "torch", "totem pole", "tow truck", "toy store",
+                        "tractor", "semi-trailer truck", "tray", "trench coat", "tricycle", "trimaran", "tripod",
+                        "triumphal arch", "trolleybus", "trombone", "hot tub", "turnstile", "typewriter keyboard",
+                        "umbrella", "unicycle", "upright piano", "vacuum cleaner", "vase", "vaulted or arched ceiling",
+                        "velvet fabric", "vending machine", "vestment", "viaduct", "violin", "volleyball",
+                        "waffle iron", "wall clock", "wallet", "wardrobe", "military aircraft", "sink",
+                        "washing machine", "water bottle", "water jug", "water tower", "whiskey jug", "whistle",
+                        "hair wig", "window screen", "window shade", "Windsor tie", "wine bottle", "airplane wing",
+                        "wok", "wooden spoon", "wool", "split-rail fence", "shipwreck", "sailboat", "yurt", "website",
+                        "comic book", "crossword", "traffic or street sign", "traffic light", "dust jacket", "menu",
+                        "plate", "guacamole", "consomme", "hot pot", "trifle", "ice cream", "popsicle", "baguette",
+                        "bagel", "pretzel", "cheeseburger", "hot dog", "mashed potatoes", "cabbage", "broccoli",
+                        "cauliflower", "zucchini", "spaghetti squash", "acorn squash", "butternut squash", "cucumber",
+                        "artichoke", "bell pepper", "cardoon", "mushroom", "Granny Smith apple", "strawberry", "orange",
+                        "lemon", "fig", "pineapple", "banana", "jackfruit", "cherimoya (custard apple)", "pomegranate",
+                        "hay", "carbonara", "chocolate syrup", "dough", "meatloaf", "pizza", "pot pie", "burrito",
+                        "red wine", "espresso", "tea cup", "eggnog", "mountain", "bubble", "cliff", "coral reef",
+                        "geyser", "lakeshore", "promontory", "sandbar", "beach", "valley", "volcano", "baseball player",
+                        "bridegroom", "scuba diver", "rapeseed", "daisy", "yellow lady's slipper", "corn", "acorn",
+                        "rose hip", "horse chestnut seed", "coral fungus", "agaric", "gyromitra", "stinkhorn mushroom",
+                        "earth star fungus", "hen of the woods mushroom", "bolete", "corn cob", "toilet paper"]
+
+    # cifar100_classes = [name for name in os.listdir(args.data_path)
+    #                 if os.path.isdir(os.path.join(args.data_path, name))]
+
+    imagenet_single_template = [
+        'a photo of a {}.',
+    ]
+
+    imagenet_7_templates = [
+        'itap of a {}.',
+        'a origami {}.',
+        'a bad photo of the {}.',
+        'a photo of the large {}.',
+        'a {} in a video game.',
+        'art of the {}.',
+        'a photo of the small {}.',
+    ]
+
+
+    print('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    print('load data')
+    valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    print('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, imagenet_classes, imagenet_7_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    print(f'accuracy with text proxy: {top1:.2f}')
+
+    print('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+
+    from collections import defaultdict
+    class_freq = torch.zeros(num_class).cuda()
+
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+            beta = args.beta * math.sqrt((i + 1) / n)
+
+            x = image_feat[idx[i], :]
+
+            vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+
+            pred_class = torch.argmax(vision_label).item()
+            class_freq[pred_class] += 1
+
+            # Step 3: compute freq-based pi_t / class_freq.sum() 肯定大于 0
+            freq_prior = class_freq / class_freq.sum() if class_freq.sum() > 0 else torch.ones_like(class_freq) / num_class
+            pi_t = (1-adpt_weight) * (args.alpha / num_class) + adpt_weight * freq_prior
+
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+
+            rho -= rlr * (tlabel - pi_t)
+            rho[rho < 0] = 0
+
+            text_label[i, :] = tlabel
+            # vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            grad = torch.outer(x, vision_label - tlabel)
+            w -= (lr / args.tau_i) * grad
+            w = F.normalize(w, dim=0)
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+
+        # MAPLS - EM Algorithm
+        pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+        qy = mapls(combo_label, pz = pz, qy_mode = "soft", max_iter = 100, lam = 0.8)
+
+        w = np.array(qy) / np.array(pz)
+        if combo_label.is_cuda:
+            combo_label_cpu = combo_label.cpu()
+        qy_probs = lsc(combo_label_cpu, 1.0/w)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+    print('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    print('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    print('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    # lams = [1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1]
+    adpt_weights = [0.0, 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4]
+    for adpt_weight in adpt_weights:
+        main(adpt_weight)
+

OnZeta/main_online_imagenet_inloop_online_MAPLS_only.py

@@ -0,0 +1,326 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+import logging
+from datetime import datetime
+
+log_filename = os.path.join("logs", f"debug_onzeta_eval_{datetime.now().strftime('%Y-%m-%d_%H-%M-%S')}.log")
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(message)s',
+    handlers=[
+        logging.FileHandler(log_filename),
+        logging.StreamHandler()
+    ]
+)
+
+from MAPLS.mapls_cuda import mapls_torch
+from MAPLS.common_cuda import lsc_torch
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='/home/li325/space_mlai/pengxiao_space/dataset/ImageNet/', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='RN50',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=256, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.8, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+
+# parser.add_argument('--lam', default=0.6, type=float)
+
+
+def main(lam):
+
+    args = parser.parse_args()
+    print(args)
+    imagenet_classes = ["tench", "goldfish", "great white shark", "tiger shark", "hammerhead shark", "electric ray",
+                        "stingray", "rooster", "hen", "ostrich", "brambling", "goldfinch", "house finch", "junco",
+                        "indigo bunting", "American robin", "bulbul", "jay", "magpie", "chickadee", "American dipper",
+                        "kite (bird of prey)", "bald eagle", "vulture", "great grey owl", "fire salamander",
+                        "smooth newt", "newt", "spotted salamander", "axolotl", "American bullfrog", "tree frog",
+                        "tailed frog", "loggerhead sea turtle", "leatherback sea turtle", "mud turtle", "terrapin",
+                        "box turtle", "banded gecko", "green iguana", "Carolina anole",
+                        "desert grassland whiptail lizard", "agama", "frilled-necked lizard", "alligator lizard",
+                        "Gila monster", "European green lizard", "chameleon", "Komodo dragon", "Nile crocodile",
+                        "American alligator", "triceratops", "worm snake", "ring-necked snake",
+                        "eastern hog-nosed snake", "smooth green snake", "kingsnake", "garter snake", "water snake",
+                        "vine snake", "night snake", "boa constrictor", "African rock python", "Indian cobra",
+                        "green mamba", "sea snake", "Saharan horned viper", "eastern diamondback rattlesnake",
+                        "sidewinder rattlesnake", "trilobite", "harvestman", "scorpion", "yellow garden spider",
+                        "barn spider", "European garden spider", "southern black widow", "tarantula", "wolf spider",
+                        "tick", "centipede", "black grouse", "ptarmigan", "ruffed grouse", "prairie grouse", "peafowl",
+                        "quail", "partridge", "african grey parrot", "macaw", "sulphur-crested cockatoo", "lorikeet",
+                        "coucal", "bee eater", "hornbill", "hummingbird", "jacamar", "toucan", "duck",
+                        "red-breasted merganser", "goose", "black swan", "tusker", "echidna", "platypus", "wallaby",
+                        "koala", "wombat", "jellyfish", "sea anemone", "brain coral", "flatworm", "nematode", "conch",
+                        "snail", "slug", "sea slug", "chiton", "chambered nautilus", "Dungeness crab", "rock crab",
+                        "fiddler crab", "red king crab", "American lobster", "spiny lobster", "crayfish", "hermit crab",
+                        "isopod", "white stork", "black stork", "spoonbill", "flamingo", "little blue heron",
+                        "great egret", "bittern bird", "crane bird", "limpkin", "common gallinule", "American coot",
+                        "bustard", "ruddy turnstone", "dunlin", "common redshank", "dowitcher", "oystercatcher",
+                        "pelican", "king penguin", "albatross", "grey whale", "killer whale", "dugong", "sea lion",
+                        "Chihuahua", "Japanese Chin", "Maltese", "Pekingese", "Shih Tzu", "King Charles Spaniel",
+                        "Papillon", "toy terrier", "Rhodesian Ridgeback", "Afghan Hound", "Basset Hound", "Beagle",
+                        "Bloodhound", "Bluetick Coonhound", "Black and Tan Coonhound", "Treeing Walker Coonhound",
+                        "English foxhound", "Redbone Coonhound", "borzoi", "Irish Wolfhound", "Italian Greyhound",
+                        "Whippet", "Ibizan Hound", "Norwegian Elkhound", "Otterhound", "Saluki", "Scottish Deerhound",
+                        "Weimaraner", "Staffordshire Bull Terrier", "American Staffordshire Terrier",
+                        "Bedlington Terrier", "Border Terrier", "Kerry Blue Terrier", "Irish Terrier",
+                        "Norfolk Terrier", "Norwich Terrier", "Yorkshire Terrier", "Wire Fox Terrier",
+                        "Lakeland Terrier", "Sealyham Terrier", "Airedale Terrier", "Cairn Terrier",
+                        "Australian Terrier", "Dandie Dinmont Terrier", "Boston Terrier", "Miniature Schnauzer",
+                        "Giant Schnauzer", "Standard Schnauzer", "Scottish Terrier", "Tibetan Terrier",
+                        "Australian Silky Terrier", "Soft-coated Wheaten Terrier", "West Highland White Terrier",
+                        "Lhasa Apso", "Flat-Coated Retriever", "Curly-coated Retriever", "Golden Retriever",
+                        "Labrador Retriever", "Chesapeake Bay Retriever", "German Shorthaired Pointer", "Vizsla",
+                        "English Setter", "Irish Setter", "Gordon Setter", "Brittany dog", "Clumber Spaniel",
+                        "English Springer Spaniel", "Welsh Springer Spaniel", "Cocker Spaniel", "Sussex Spaniel",
+                        "Irish Water Spaniel", "Kuvasz", "Schipperke", "Groenendael dog", "Malinois", "Briard",
+                        "Australian Kelpie", "Komondor", "Old English Sheepdog", "Shetland Sheepdog", "collie",
+                        "Border Collie", "Bouvier des Flandres dog", "Rottweiler", "German Shepherd Dog", "Dobermann",
+                        "Miniature Pinscher", "Greater Swiss Mountain Dog", "Bernese Mountain Dog",
+                        "Appenzeller Sennenhund", "Entlebucher Sennenhund", "Boxer", "Bullmastiff", "Tibetan Mastiff",
+                        "French Bulldog", "Great Dane", "St. Bernard", "husky", "Alaskan Malamute", "Siberian Husky",
+                        "Dalmatian", "Affenpinscher", "Basenji", "pug", "Leonberger", "Newfoundland dog",
+                        "Great Pyrenees dog", "Samoyed", "Pomeranian", "Chow Chow", "Keeshond", "brussels griffon",
+                        "Pembroke Welsh Corgi", "Cardigan Welsh Corgi", "Toy Poodle", "Miniature Poodle",
+                        "Standard Poodle", "Mexican hairless dog (xoloitzcuintli)", "grey wolf", "Alaskan tundra wolf",
+                        "red wolf or maned wolf", "coyote", "dingo", "dhole", "African wild dog", "hyena", "red fox",
+                        "kit fox", "Arctic fox", "grey fox", "tabby cat", "tiger cat", "Persian cat", "Siamese cat",
+                        "Egyptian Mau", "cougar", "lynx", "leopard", "snow leopard", "jaguar", "lion", "tiger",
+                        "cheetah", "brown bear", "American black bear", "polar bear", "sloth bear", "mongoose",
+                        "meerkat", "tiger beetle", "ladybug", "ground beetle", "longhorn beetle", "leaf beetle",
+                        "dung beetle", "rhinoceros beetle", "weevil", "fly", "bee", "ant", "grasshopper",
+                        "cricket insect", "stick insect", "cockroach", "praying mantis", "cicada", "leafhopper",
+                        "lacewing", "dragonfly", "damselfly", "red admiral butterfly", "ringlet butterfly",
+                        "monarch butterfly", "small white butterfly", "sulphur butterfly", "gossamer-winged butterfly",
+                        "starfish", "sea urchin", "sea cucumber", "cottontail rabbit", "hare", "Angora rabbit",
+                        "hamster", "porcupine", "fox squirrel", "marmot", "beaver", "guinea pig", "common sorrel horse",
+                        "zebra", "pig", "wild boar", "warthog", "hippopotamus", "ox", "water buffalo", "bison",
+                        "ram (adult male sheep)", "bighorn sheep", "Alpine ibex", "hartebeest", "impala (antelope)",
+                        "gazelle", "arabian camel", "llama", "weasel", "mink", "European polecat",
+                        "black-footed ferret", "otter", "skunk", "badger", "armadillo", "three-toed sloth", "orangutan",
+                        "gorilla", "chimpanzee", "gibbon", "siamang", "guenon", "patas monkey", "baboon", "macaque",
+                        "langur", "black-and-white colobus", "proboscis monkey", "marmoset", "white-headed capuchin",
+                        "howler monkey", "titi monkey", "Geoffroy's spider monkey", "common squirrel monkey",
+                        "ring-tailed lemur", "indri", "Asian elephant", "African bush elephant", "red panda",
+                        "giant panda", "snoek fish", "eel", "silver salmon", "rock beauty fish", "clownfish",
+                        "sturgeon", "gar fish", "lionfish", "pufferfish", "abacus", "abaya", "academic gown",
+                        "accordion", "acoustic guitar", "aircraft carrier", "airliner", "airship", "altar", "ambulance",
+                        "amphibious vehicle", "analog clock", "apiary", "apron", "trash can", "assault rifle",
+                        "backpack", "bakery", "balance beam", "balloon", "ballpoint pen", "Band-Aid", "banjo",
+                        "baluster / handrail", "barbell", "barber chair", "barbershop", "barn", "barometer", "barrel",
+                        "wheelbarrow", "baseball", "basketball", "bassinet", "bassoon", "swimming cap", "bath towel",
+                        "bathtub", "station wagon", "lighthouse", "beaker", "military hat (bearskin or shako)",
+                        "beer bottle", "beer glass", "bell tower", "baby bib", "tandem bicycle", "bikini",
+                        "ring binder", "binoculars", "birdhouse", "boathouse", "bobsleigh", "bolo tie", "poke bonnet",
+                        "bookcase", "bookstore", "bottle cap", "hunting bow", "bow tie", "brass memorial plaque", "bra",
+                        "breakwater", "breastplate", "broom", "bucket", "buckle", "bulletproof vest",
+                        "high-speed train", "butcher shop", "taxicab", "cauldron", "candle", "cannon", "canoe",
+                        "can opener", "cardigan", "car mirror", "carousel", "tool kit", "cardboard box / carton",
+                        "car wheel", "automated teller machine", "cassette", "cassette player", "castle", "catamaran",
+                        "CD player", "cello", "mobile phone", "chain", "chain-link fence", "chain mail", "chainsaw",
+                        "storage chest", "chiffonier", "bell or wind chime", "china cabinet", "Christmas stocking",
+                        "church", "movie theater", "cleaver", "cliff dwelling", "cloak", "clogs", "cocktail shaker",
+                        "coffee mug", "coffeemaker", "spiral or coil", "combination lock", "computer keyboard",
+                        "candy store", "container ship", "convertible", "corkscrew", "cornet", "cowboy boot",
+                        "cowboy hat", "cradle", "construction crane", "crash helmet", "crate", "infant bed",
+                        "Crock Pot", "croquet ball", "crutch", "cuirass", "dam", "desk", "desktop computer",
+                        "rotary dial telephone", "diaper", "digital clock", "digital watch", "dining table",
+                        "dishcloth", "dishwasher", "disc brake", "dock", "dog sled", "dome", "doormat", "drilling rig",
+                        "drum", "drumstick", "dumbbell", "Dutch oven", "electric fan", "electric guitar",
+                        "electric locomotive", "entertainment center", "envelope", "espresso machine", "face powder",
+                        "feather boa", "filing cabinet", "fireboat", "fire truck", "fire screen", "flagpole", "flute",
+                        "folding chair", "football helmet", "forklift", "fountain", "fountain pen", "four-poster bed",
+                        "freight car", "French horn", "frying pan", "fur coat", "garbage truck",
+                        "gas mask or respirator", "gas pump", "goblet", "go-kart", "golf ball", "golf cart", "gondola",
+                        "gong", "gown", "grand piano", "greenhouse", "radiator grille", "grocery store", "guillotine",
+                        "hair clip", "hair spray", "half-track", "hammer", "hamper", "hair dryer", "hand-held computer",
+                        "handkerchief", "hard disk drive", "harmonica", "harp", "combine harvester", "hatchet",
+                        "holster", "home theater", "honeycomb", "hook", "hoop skirt", "gymnastic horizontal bar",
+                        "horse-drawn vehicle", "hourglass", "iPod", "clothes iron", "carved pumpkin", "jeans", "jeep",
+                        "T-shirt", "jigsaw puzzle", "rickshaw", "joystick", "kimono", "knee pad", "knot", "lab coat",
+                        "ladle", "lampshade", "laptop computer", "lawn mower", "lens cap", "letter opener", "library",
+                        "lifeboat", "lighter", "limousine", "ocean liner", "lipstick", "slip-on shoe", "lotion",
+                        "music speaker", "loupe magnifying glass", "sawmill", "magnetic compass", "messenger bag",
+                        "mailbox", "tights", "one-piece bathing suit", "manhole cover", "maraca", "marimba", "mask",
+                        "matchstick", "maypole", "maze", "measuring cup", "medicine cabinet", "megalith", "microphone",
+                        "microwave oven", "military uniform", "milk can", "minibus", "miniskirt", "minivan", "missile",
+                        "mitten", "mixing bowl", "mobile home", "ford model t", "modem", "monastery", "monitor",
+                        "moped", "mortar and pestle", "graduation cap", "mosque", "mosquito net", "vespa",
+                        "mountain bike", "tent", "computer mouse", "mousetrap", "moving van", "muzzle", "metal nail",
+                        "neck brace", "necklace", "baby pacifier", "notebook computer", "obelisk", "oboe", "ocarina",
+                        "odometer", "oil filter", "pipe organ", "oscilloscope", "overskirt", "bullock cart",
+                        "oxygen mask", "product packet / packaging", "paddle", "paddle wheel", "padlock", "paintbrush",
+                        "pajamas", "palace", "pan flute", "paper towel", "parachute", "parallel bars", "park bench",
+                        "parking meter", "railroad car", "patio", "payphone", "pedestal", "pencil case",
+                        "pencil sharpener", "perfume", "Petri dish", "photocopier", "plectrum", "Pickelhaube",
+                        "picket fence", "pickup truck", "pier", "piggy bank", "pill bottle", "pillow", "ping-pong ball",
+                        "pinwheel", "pirate ship", "drink pitcher", "block plane", "planetarium", "plastic bag",
+                        "plate rack", "farm plow", "plunger", "Polaroid camera", "pole", "police van", "poncho",
+                        "pool table", "soda bottle", "plant pot", "potter's wheel", "power drill", "prayer rug",
+                        "printer", "prison", "missile", "projector", "hockey puck", "punching bag", "purse", "quill",
+                        "quilt", "race car", "racket", "radiator", "radio", "radio telescope", "rain barrel",
+                        "recreational vehicle", "fishing casting reel", "reflex camera", "refrigerator",
+                        "remote control", "restaurant", "revolver", "rifle", "rocking chair", "rotisserie", "eraser",
+                        "rugby ball", "ruler measuring stick", "sneaker", "safe", "safety pin", "salt shaker", "sandal",
+                        "sarong", "saxophone", "scabbard", "weighing scale", "school bus", "schooner", "scoreboard",
+                        "CRT monitor", "screw", "screwdriver", "seat belt", "sewing machine", "shield", "shoe store",
+                        "shoji screen / room divider", "shopping basket", "shopping cart", "shovel", "shower cap",
+                        "shower curtain", "ski", "balaclava ski mask", "sleeping bag", "slide rule", "sliding door",
+                        "slot machine", "snorkel", "snowmobile", "snowplow", "soap dispenser", "soccer ball", "sock",
+                        "solar thermal collector", "sombrero", "soup bowl", "keyboard space bar", "space heater",
+                        "space shuttle", "spatula", "motorboat", "spider web", "spindle", "sports car", "spotlight",
+                        "stage", "steam locomotive", "through arch bridge", "steel drum", "stethoscope", "scarf",
+                        "stone wall", "stopwatch", "stove", "strainer", "tram", "stretcher", "couch", "stupa",
+                        "submarine", "suit", "sundial", "sunglasses", "sunglasses", "sunscreen", "suspension bridge",
+                        "mop", "sweatshirt", "swim trunks / shorts", "swing", "electrical switch", "syringe",
+                        "table lamp", "tank", "tape player", "teapot", "teddy bear", "television", "tennis ball",
+                        "thatched roof", "front curtain", "thimble", "threshing machine", "throne", "tile roof",
+                        "toaster", "tobacco shop", "toilet seat", "torch", "totem pole", "tow truck", "toy store",
+                        "tractor", "semi-trailer truck", "tray", "trench coat", "tricycle", "trimaran", "tripod",
+                        "triumphal arch", "trolleybus", "trombone", "hot tub", "turnstile", "typewriter keyboard",
+                        "umbrella", "unicycle", "upright piano", "vacuum cleaner", "vase", "vaulted or arched ceiling",
+                        "velvet fabric", "vending machine", "vestment", "viaduct", "violin", "volleyball",
+                        "waffle iron", "wall clock", "wallet", "wardrobe", "military aircraft", "sink",
+                        "washing machine", "water bottle", "water jug", "water tower", "whiskey jug", "whistle",
+                        "hair wig", "window screen", "window shade", "Windsor tie", "wine bottle", "airplane wing",
+                        "wok", "wooden spoon", "wool", "split-rail fence", "shipwreck", "sailboat", "yurt", "website",
+                        "comic book", "crossword", "traffic or street sign", "traffic light", "dust jacket", "menu",
+                        "plate", "guacamole", "consomme", "hot pot", "trifle", "ice cream", "popsicle", "baguette",
+                        "bagel", "pretzel", "cheeseburger", "hot dog", "mashed potatoes", "cabbage", "broccoli",
+                        "cauliflower", "zucchini", "spaghetti squash", "acorn squash", "butternut squash", "cucumber",
+                        "artichoke", "bell pepper", "cardoon", "mushroom", "Granny Smith apple", "strawberry", "orange",
+                        "lemon", "fig", "pineapple", "banana", "jackfruit", "cherimoya (custard apple)", "pomegranate",
+                        "hay", "carbonara", "chocolate syrup", "dough", "meatloaf", "pizza", "pot pie", "burrito",
+                        "red wine", "espresso", "tea cup", "eggnog", "mountain", "bubble", "cliff", "coral reef",
+                        "geyser", "lakeshore", "promontory", "sandbar", "beach", "valley", "volcano", "baseball player",
+                        "bridegroom", "scuba diver", "rapeseed", "daisy", "yellow lady's slipper", "corn", "acorn",
+                        "rose hip", "horse chestnut seed", "coral fungus", "agaric", "gyromitra", "stinkhorn mushroom",
+                        "earth star fungus", "hen of the woods mushroom", "bolete", "corn cob", "toilet paper"]
+
+    # cifar100_classes = [name for name in os.listdir(args.data_path)
+    #                 if os.path.isdir(os.path.join(args.data_path, name))]
+
+    # imagenet_7_templates = [
+    #     'a photo of a {}.',
+    # ]
+
+    imagenet_7_templates = [
+        'itap of a {}.',
+        'a origami {}.',
+        'a bad photo of the {}.',
+        'a photo of the large {}.',
+        'a {} in a video game.',
+        'art of the {}.',
+        'a photo of the small {}.',
+    ]
+
+    print('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    print('load data')
+    valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    print('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, imagenet_classes, imagenet_7_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    print(f'accuracy with text proxy: {top1:.2f}')
+
+    print('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        for i in range(n):
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            text_label[i, :] = tlabel
+            ####################################################
+            # MAPLS - EM Algorithm
+            pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+            pz = torch.tensor(pz, dtype=torch.float32, device='cuda')
+            qy = mapls_torch(text_label, pz=pz, qy_mode="soft", max_iter=50, lam=lam)
+            w_mapls = qy.to("cuda") / pz.to("cuda")
+            qy_probs = lsc_torch(text_label, 1.0 / w_mapls)
+            # MAPLS - EM Algorithm
+            ####################################################
+        qy_probs = torch.from_numpy(qy_probs.cpu().numpy()).float()
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+        acc_ls[iter] = (acc1_ls / n) * 100
+    logging.info('mean acc of MAPLS only in-loop with lambda {:.2f} is: {:.2f}'.format(lam, torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    lams = [0.95, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1]
+    for lam in lams:
+        main(lam)
+

OnZeta/main_online_imagenet_mapls.py

@@ -0,0 +1,417 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='/home/li325/space_mlai/pengxiao_space/dataset/ImageNet/', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='RN50',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=1, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.8, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+
+
+def main(lam):
+
+    args = parser.parse_args()
+    print(args)
+    imagenet_classes = ["tench", "goldfish", "great white shark", "tiger shark", "hammerhead shark", "electric ray",
+                        "stingray", "rooster", "hen", "ostrich", "brambling", "goldfinch", "house finch", "junco",
+                        "indigo bunting", "American robin", "bulbul", "jay", "magpie", "chickadee", "American dipper",
+                        "kite (bird of prey)", "bald eagle", "vulture", "great grey owl", "fire salamander",
+                        "smooth newt", "newt", "spotted salamander", "axolotl", "American bullfrog", "tree frog",
+                        "tailed frog", "loggerhead sea turtle", "leatherback sea turtle", "mud turtle", "terrapin",
+                        "box turtle", "banded gecko", "green iguana", "Carolina anole",
+                        "desert grassland whiptail lizard", "agama", "frilled-necked lizard", "alligator lizard",
+                        "Gila monster", "European green lizard", "chameleon", "Komodo dragon", "Nile crocodile",
+                        "American alligator", "triceratops", "worm snake", "ring-necked snake",
+                        "eastern hog-nosed snake", "smooth green snake", "kingsnake", "garter snake", "water snake",
+                        "vine snake", "night snake", "boa constrictor", "African rock python", "Indian cobra",
+                        "green mamba", "sea snake", "Saharan horned viper", "eastern diamondback rattlesnake",
+                        "sidewinder rattlesnake", "trilobite", "harvestman", "scorpion", "yellow garden spider",
+                        "barn spider", "European garden spider", "southern black widow", "tarantula", "wolf spider",
+                        "tick", "centipede", "black grouse", "ptarmigan", "ruffed grouse", "prairie grouse", "peafowl",
+                        "quail", "partridge", "african grey parrot", "macaw", "sulphur-crested cockatoo", "lorikeet",
+                        "coucal", "bee eater", "hornbill", "hummingbird", "jacamar", "toucan", "duck",
+                        "red-breasted merganser", "goose", "black swan", "tusker", "echidna", "platypus", "wallaby",
+                        "koala", "wombat", "jellyfish", "sea anemone", "brain coral", "flatworm", "nematode", "conch",
+                        "snail", "slug", "sea slug", "chiton", "chambered nautilus", "Dungeness crab", "rock crab",
+                        "fiddler crab", "red king crab", "American lobster", "spiny lobster", "crayfish", "hermit crab",
+                        "isopod", "white stork", "black stork", "spoonbill", "flamingo", "little blue heron",
+                        "great egret", "bittern bird", "crane bird", "limpkin", "common gallinule", "American coot",
+                        "bustard", "ruddy turnstone", "dunlin", "common redshank", "dowitcher", "oystercatcher",
+                        "pelican", "king penguin", "albatross", "grey whale", "killer whale", "dugong", "sea lion",
+                        "Chihuahua", "Japanese Chin", "Maltese", "Pekingese", "Shih Tzu", "King Charles Spaniel",
+                        "Papillon", "toy terrier", "Rhodesian Ridgeback", "Afghan Hound", "Basset Hound", "Beagle",
+                        "Bloodhound", "Bluetick Coonhound", "Black and Tan Coonhound", "Treeing Walker Coonhound",
+                        "English foxhound", "Redbone Coonhound", "borzoi", "Irish Wolfhound", "Italian Greyhound",
+                        "Whippet", "Ibizan Hound", "Norwegian Elkhound", "Otterhound", "Saluki", "Scottish Deerhound",
+                        "Weimaraner", "Staffordshire Bull Terrier", "American Staffordshire Terrier",
+                        "Bedlington Terrier", "Border Terrier", "Kerry Blue Terrier", "Irish Terrier",
+                        "Norfolk Terrier", "Norwich Terrier", "Yorkshire Terrier", "Wire Fox Terrier",
+                        "Lakeland Terrier", "Sealyham Terrier", "Airedale Terrier", "Cairn Terrier",
+                        "Australian Terrier", "Dandie Dinmont Terrier", "Boston Terrier", "Miniature Schnauzer",
+                        "Giant Schnauzer", "Standard Schnauzer", "Scottish Terrier", "Tibetan Terrier",
+                        "Australian Silky Terrier", "Soft-coated Wheaten Terrier", "West Highland White Terrier",
+                        "Lhasa Apso", "Flat-Coated Retriever", "Curly-coated Retriever", "Golden Retriever",
+                        "Labrador Retriever", "Chesapeake Bay Retriever", "German Shorthaired Pointer", "Vizsla",
+                        "English Setter", "Irish Setter", "Gordon Setter", "Brittany dog", "Clumber Spaniel",
+                        "English Springer Spaniel", "Welsh Springer Spaniel", "Cocker Spaniel", "Sussex Spaniel",
+                        "Irish Water Spaniel", "Kuvasz", "Schipperke", "Groenendael dog", "Malinois", "Briard",
+                        "Australian Kelpie", "Komondor", "Old English Sheepdog", "Shetland Sheepdog", "collie",
+                        "Border Collie", "Bouvier des Flandres dog", "Rottweiler", "German Shepherd Dog", "Dobermann",
+                        "Miniature Pinscher", "Greater Swiss Mountain Dog", "Bernese Mountain Dog",
+                        "Appenzeller Sennenhund", "Entlebucher Sennenhund", "Boxer", "Bullmastiff", "Tibetan Mastiff",
+                        "French Bulldog", "Great Dane", "St. Bernard", "husky", "Alaskan Malamute", "Siberian Husky",
+                        "Dalmatian", "Affenpinscher", "Basenji", "pug", "Leonberger", "Newfoundland dog",
+                        "Great Pyrenees dog", "Samoyed", "Pomeranian", "Chow Chow", "Keeshond", "brussels griffon",
+                        "Pembroke Welsh Corgi", "Cardigan Welsh Corgi", "Toy Poodle", "Miniature Poodle",
+                        "Standard Poodle", "Mexican hairless dog (xoloitzcuintli)", "grey wolf", "Alaskan tundra wolf",
+                        "red wolf or maned wolf", "coyote", "dingo", "dhole", "African wild dog", "hyena", "red fox",
+                        "kit fox", "Arctic fox", "grey fox", "tabby cat", "tiger cat", "Persian cat", "Siamese cat",
+                        "Egyptian Mau", "cougar", "lynx", "leopard", "snow leopard", "jaguar", "lion", "tiger",
+                        "cheetah", "brown bear", "American black bear", "polar bear", "sloth bear", "mongoose",
+                        "meerkat", "tiger beetle", "ladybug", "ground beetle", "longhorn beetle", "leaf beetle",
+                        "dung beetle", "rhinoceros beetle", "weevil", "fly", "bee", "ant", "grasshopper",
+                        "cricket insect", "stick insect", "cockroach", "praying mantis", "cicada", "leafhopper",
+                        "lacewing", "dragonfly", "damselfly", "red admiral butterfly", "ringlet butterfly",
+                        "monarch butterfly", "small white butterfly", "sulphur butterfly", "gossamer-winged butterfly",
+                        "starfish", "sea urchin", "sea cucumber", "cottontail rabbit", "hare", "Angora rabbit",
+                        "hamster", "porcupine", "fox squirrel", "marmot", "beaver", "guinea pig", "common sorrel horse",
+                        "zebra", "pig", "wild boar", "warthog", "hippopotamus", "ox", "water buffalo", "bison",
+                        "ram (adult male sheep)", "bighorn sheep", "Alpine ibex", "hartebeest", "impala (antelope)",
+                        "gazelle", "arabian camel", "llama", "weasel", "mink", "European polecat",
+                        "black-footed ferret", "otter", "skunk", "badger", "armadillo", "three-toed sloth", "orangutan",
+                        "gorilla", "chimpanzee", "gibbon", "siamang", "guenon", "patas monkey", "baboon", "macaque",
+                        "langur", "black-and-white colobus", "proboscis monkey", "marmoset", "white-headed capuchin",
+                        "howler monkey", "titi monkey", "Geoffroy's spider monkey", "common squirrel monkey",
+                        "ring-tailed lemur", "indri", "Asian elephant", "African bush elephant", "red panda",
+                        "giant panda", "snoek fish", "eel", "silver salmon", "rock beauty fish", "clownfish",
+                        "sturgeon", "gar fish", "lionfish", "pufferfish", "abacus", "abaya", "academic gown",
+                        "accordion", "acoustic guitar", "aircraft carrier", "airliner", "airship", "altar", "ambulance",
+                        "amphibious vehicle", "analog clock", "apiary", "apron", "trash can", "assault rifle",
+                        "backpack", "bakery", "balance beam", "balloon", "ballpoint pen", "Band-Aid", "banjo",
+                        "baluster / handrail", "barbell", "barber chair", "barbershop", "barn", "barometer", "barrel",
+                        "wheelbarrow", "baseball", "basketball", "bassinet", "bassoon", "swimming cap", "bath towel",
+                        "bathtub", "station wagon", "lighthouse", "beaker", "military hat (bearskin or shako)",
+                        "beer bottle", "beer glass", "bell tower", "baby bib", "tandem bicycle", "bikini",
+                        "ring binder", "binoculars", "birdhouse", "boathouse", "bobsleigh", "bolo tie", "poke bonnet",
+                        "bookcase", "bookstore", "bottle cap", "hunting bow", "bow tie", "brass memorial plaque", "bra",
+                        "breakwater", "breastplate", "broom", "bucket", "buckle", "bulletproof vest",
+                        "high-speed train", "butcher shop", "taxicab", "cauldron", "candle", "cannon", "canoe",
+                        "can opener", "cardigan", "car mirror", "carousel", "tool kit", "cardboard box / carton",
+                        "car wheel", "automated teller machine", "cassette", "cassette player", "castle", "catamaran",
+                        "CD player", "cello", "mobile phone", "chain", "chain-link fence", "chain mail", "chainsaw",
+                        "storage chest", "chiffonier", "bell or wind chime", "china cabinet", "Christmas stocking",
+                        "church", "movie theater", "cleaver", "cliff dwelling", "cloak", "clogs", "cocktail shaker",
+                        "coffee mug", "coffeemaker", "spiral or coil", "combination lock", "computer keyboard",
+                        "candy store", "container ship", "convertible", "corkscrew", "cornet", "cowboy boot",
+                        "cowboy hat", "cradle", "construction crane", "crash helmet", "crate", "infant bed",
+                        "Crock Pot", "croquet ball", "crutch", "cuirass", "dam", "desk", "desktop computer",
+                        "rotary dial telephone", "diaper", "digital clock", "digital watch", "dining table",
+                        "dishcloth", "dishwasher", "disc brake", "dock", "dog sled", "dome", "doormat", "drilling rig",
+                        "drum", "drumstick", "dumbbell", "Dutch oven", "electric fan", "electric guitar",
+                        "electric locomotive", "entertainment center", "envelope", "espresso machine", "face powder",
+                        "feather boa", "filing cabinet", "fireboat", "fire truck", "fire screen", "flagpole", "flute",
+                        "folding chair", "football helmet", "forklift", "fountain", "fountain pen", "four-poster bed",
+                        "freight car", "French horn", "frying pan", "fur coat", "garbage truck",
+                        "gas mask or respirator", "gas pump", "goblet", "go-kart", "golf ball", "golf cart", "gondola",
+                        "gong", "gown", "grand piano", "greenhouse", "radiator grille", "grocery store", "guillotine",
+                        "hair clip", "hair spray", "half-track", "hammer", "hamper", "hair dryer", "hand-held computer",
+                        "handkerchief", "hard disk drive", "harmonica", "harp", "combine harvester", "hatchet",
+                        "holster", "home theater", "honeycomb", "hook", "hoop skirt", "gymnastic horizontal bar",
+                        "horse-drawn vehicle", "hourglass", "iPod", "clothes iron", "carved pumpkin", "jeans", "jeep",
+                        "T-shirt", "jigsaw puzzle", "rickshaw", "joystick", "kimono", "knee pad", "knot", "lab coat",
+                        "ladle", "lampshade", "laptop computer", "lawn mower", "lens cap", "letter opener", "library",
+                        "lifeboat", "lighter", "limousine", "ocean liner", "lipstick", "slip-on shoe", "lotion",
+                        "music speaker", "loupe magnifying glass", "sawmill", "magnetic compass", "messenger bag",
+                        "mailbox", "tights", "one-piece bathing suit", "manhole cover", "maraca", "marimba", "mask",
+                        "matchstick", "maypole", "maze", "measuring cup", "medicine cabinet", "megalith", "microphone",
+                        "microwave oven", "military uniform", "milk can", "minibus", "miniskirt", "minivan", "missile",
+                        "mitten", "mixing bowl", "mobile home", "ford model t", "modem", "monastery", "monitor",
+                        "moped", "mortar and pestle", "graduation cap", "mosque", "mosquito net", "vespa",
+                        "mountain bike", "tent", "computer mouse", "mousetrap", "moving van", "muzzle", "metal nail",
+                        "neck brace", "necklace", "baby pacifier", "notebook computer", "obelisk", "oboe", "ocarina",
+                        "odometer", "oil filter", "pipe organ", "oscilloscope", "overskirt", "bullock cart",
+                        "oxygen mask", "product packet / packaging", "paddle", "paddle wheel", "padlock", "paintbrush",
+                        "pajamas", "palace", "pan flute", "paper towel", "parachute", "parallel bars", "park bench",
+                        "parking meter", "railroad car", "patio", "payphone", "pedestal", "pencil case",
+                        "pencil sharpener", "perfume", "Petri dish", "photocopier", "plectrum", "Pickelhaube",
+                        "picket fence", "pickup truck", "pier", "piggy bank", "pill bottle", "pillow", "ping-pong ball",
+                        "pinwheel", "pirate ship", "drink pitcher", "block plane", "planetarium", "plastic bag",
+                        "plate rack", "farm plow", "plunger", "Polaroid camera", "pole", "police van", "poncho",
+                        "pool table", "soda bottle", "plant pot", "potter's wheel", "power drill", "prayer rug",
+                        "printer", "prison", "missile", "projector", "hockey puck", "punching bag", "purse", "quill",
+                        "quilt", "race car", "racket", "radiator", "radio", "radio telescope", "rain barrel",
+                        "recreational vehicle", "fishing casting reel", "reflex camera", "refrigerator",
+                        "remote control", "restaurant", "revolver", "rifle", "rocking chair", "rotisserie", "eraser",
+                        "rugby ball", "ruler measuring stick", "sneaker", "safe", "safety pin", "salt shaker", "sandal",
+                        "sarong", "saxophone", "scabbard", "weighing scale", "school bus", "schooner", "scoreboard",
+                        "CRT monitor", "screw", "screwdriver", "seat belt", "sewing machine", "shield", "shoe store",
+                        "shoji screen / room divider", "shopping basket", "shopping cart", "shovel", "shower cap",
+                        "shower curtain", "ski", "balaclava ski mask", "sleeping bag", "slide rule", "sliding door",
+                        "slot machine", "snorkel", "snowmobile", "snowplow", "soap dispenser", "soccer ball", "sock",
+                        "solar thermal collector", "sombrero", "soup bowl", "keyboard space bar", "space heater",
+                        "space shuttle", "spatula", "motorboat", "spider web", "spindle", "sports car", "spotlight",
+                        "stage", "steam locomotive", "through arch bridge", "steel drum", "stethoscope", "scarf",
+                        "stone wall", "stopwatch", "stove", "strainer", "tram", "stretcher", "couch", "stupa",
+                        "submarine", "suit", "sundial", "sunglasses", "sunglasses", "sunscreen", "suspension bridge",
+                        "mop", "sweatshirt", "swim trunks / shorts", "swing", "electrical switch", "syringe",
+                        "table lamp", "tank", "tape player", "teapot", "teddy bear", "television", "tennis ball",
+                        "thatched roof", "front curtain", "thimble", "threshing machine", "throne", "tile roof",
+                        "toaster", "tobacco shop", "toilet seat", "torch", "totem pole", "tow truck", "toy store",
+                        "tractor", "semi-trailer truck", "tray", "trench coat", "tricycle", "trimaran", "tripod",
+                        "triumphal arch", "trolleybus", "trombone", "hot tub", "turnstile", "typewriter keyboard",
+                        "umbrella", "unicycle", "upright piano", "vacuum cleaner", "vase", "vaulted or arched ceiling",
+                        "velvet fabric", "vending machine", "vestment", "viaduct", "violin", "volleyball",
+                        "waffle iron", "wall clock", "wallet", "wardrobe", "military aircraft", "sink",
+                        "washing machine", "water bottle", "water jug", "water tower", "whiskey jug", "whistle",
+                        "hair wig", "window screen", "window shade", "Windsor tie", "wine bottle", "airplane wing",
+                        "wok", "wooden spoon", "wool", "split-rail fence", "shipwreck", "sailboat", "yurt", "website",
+                        "comic book", "crossword", "traffic or street sign", "traffic light", "dust jacket", "menu",
+                        "plate", "guacamole", "consomme", "hot pot", "trifle", "ice cream", "popsicle", "baguette",
+                        "bagel", "pretzel", "cheeseburger", "hot dog", "mashed potatoes", "cabbage", "broccoli",
+                        "cauliflower", "zucchini", "spaghetti squash", "acorn squash", "butternut squash", "cucumber",
+                        "artichoke", "bell pepper", "cardoon", "mushroom", "Granny Smith apple", "strawberry", "orange",
+                        "lemon", "fig", "pineapple", "banana", "jackfruit", "cherimoya (custard apple)", "pomegranate",
+                        "hay", "carbonara", "chocolate syrup", "dough", "meatloaf", "pizza", "pot pie", "burrito",
+                        "red wine", "espresso", "tea cup", "eggnog", "mountain", "bubble", "cliff", "coral reef",
+                        "geyser", "lakeshore", "promontory", "sandbar", "beach", "valley", "volcano", "baseball player",
+                        "bridegroom", "scuba diver", "rapeseed", "daisy", "yellow lady's slipper", "corn", "acorn",
+                        "rose hip", "horse chestnut seed", "coral fungus", "agaric", "gyromitra", "stinkhorn mushroom",
+                        "earth star fungus", "hen of the woods mushroom", "bolete", "corn cob", "toilet paper"]
+
+    # cifar100_classes = [name for name in os.listdir(args.data_path)
+    #                 if os.path.isdir(os.path.join(args.data_path, name))]
+
+    imagenet_single_template = [
+        'a photo of a {}.',
+    ]
+
+    imagenet_7_templates = [
+        'itap of a {}.',
+        'a origami {}.',
+        'a bad photo of the {}.',
+        'a photo of the large {}.',
+        'a {} in a video game.',
+        'art of the {}.',
+        'a photo of the small {}.',
+    ]
+
+    # imagenet_7_templates = [
+    #     'a bad photo of a {}.',
+    #     'a photo of many {}.',
+    #     'a sculpture of a {}.',
+    #     'a photo of the hard to see {}.',
+    #     'a low resolution photo of the {}.',
+    #     'a rendering of a {}.',
+    #     'graffiti of a {}.',
+    #     'a bad photo of the {}.',
+    #     'a cropped photo of the {}.',
+    #     'a tattoo of a {}.',
+    #     'the embroidered {}.',
+    #     'a photo of a hard to see {}.',
+    #     'a bright photo of a {}.',
+    #     'a photo of a clean {}.',
+    #     'a photo of a dirty {}.',
+    #     'a dark photo of the {}.',
+    #     'a drawing of a {}.',
+    #     'a photo of my {}.',
+    #     'the plastic {}.',
+    #     'a photo of the cool {}.',
+    #     'a close-up photo of a {}.',
+    #     'a black and white photo of the {}.',
+    #     'a painting of the {}.',
+    #     'a painting of a {}.',
+    #     'a pixelated photo of the {}.',
+    #     'a sculpture of the {}.',
+    #     'a bright photo of the {}.',
+    #     'a cropped photo of a {}.',
+    #     'a plastic {}.',
+    #     'a photo of the dirty {}.',
+    #     'a jpeg corrupted photo of a {}.',
+    #     'a blurry photo of the {}.',
+    #     'a photo of the {}.',
+    #     'a good photo of the {}.',
+    #     'a rendering of the {}.',
+    #     'a {} in a video game.',
+    #     'a photo of one {}.',
+    #     'a doodle of a {}.',
+    #     'a close-up photo of the {}.',
+    #     'a photo of a {}.',
+    #     'the origami {}.',
+    #     'the {} in a video game.',
+    #     'a sketch of a {}.',
+    #     'a doodle of the {}.',
+    #     'a origami {}.',
+    #     'a low resolution photo of a {}.',
+    #     'the toy {}.',
+    #     'a rendition of the {}.',
+    #     'a photo of the clean {}.',
+    #     'a photo of a large {}.',
+    #     'a rendition of a {}.',
+    #     'a photo of a nice {}.',
+    #     'a photo of a weird {}.',
+    #     'a blurry photo of a {}.',
+    #     'a cartoon {}.',
+    #     'art of a {}.',
+    #     'a sketch of the {}.',
+    #     'a embroidered {}.',
+    #     'a pixelated photo of a {}.',
+    #     'itap of the {}.',
+    #     'a jpeg corrupted photo of the {}.',
+    #     'a good photo of a {}.',
+    #     'a plushie {}.',
+    #     'a photo of the nice {}.',
+    #     'a photo of the small {}.',
+    #     'a photo of the weird {}.',
+    #     'the cartoon {}.',
+    #     'art of the {}.',
+    #     'a drawing of the {}.',
+    #     'a photo of the large {}.',
+    #     'a black and white photo of a {}.',
+    #     'the plushie {}.',
+    #     'a dark photo of a {}.',
+    #     'itap of a {}.',
+    #     'graffiti of the {}.',
+    #     'a toy {}.',
+    #     'itap of my {}.',
+    #     'a photo of a cool {}.',
+    #     'a photo of a small {}.',
+    #     'a tattoo of the {}.',
+    # ]
+
+    print('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    print('load data')
+    valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    print('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, imagenet_classes, imagenet_7_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    print(f'accuracy with text proxy: {top1:.2f}')
+
+    print('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+            beta = args.beta * math.sqrt((i + 1) / n)
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+            rho -= rlr * (tlabel - args.alpha / num_class)
+            rho[rho < 0] = 0
+            text_label[i, :] = tlabel
+            vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            grad = torch.outer(x, vision_label - tlabel)
+            w -= (lr / args.tau_i) * grad
+            w = F.normalize(w, dim=0)
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+
+        # MAPLS - EM Algorithm
+        pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+        qy = mapls(combo_label, pz = pz, qy_mode = "soft", max_iter = 100, lam = lam)
+
+        w = np.array(qy) / np.array(pz)
+        if combo_label.is_cuda:
+            combo_label_cpu = combo_label.cpu()
+        qy_probs = lsc(combo_label_cpu, 1.0/w)
+        qy_probs = torch.from_numpy(qy_probs)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+    print('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    print('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    print('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    # lams = [1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1]
+    lams = [0.6]
+    for lam in lams:
+        main(lam)
+

OnZeta/main_online_imagenet_mapls_aug.py

@@ -0,0 +1,434 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+from torchvision.transforms.functional import to_pil_image
+import numpy as np
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='/home/lt_test/projects/datasets/ImageNet/', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='RN50',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=512, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.8, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+from torchvision import transforms
+
+def main(lam):
+    augmentations = [
+        transforms.Compose([
+            # transforms.RandomResizedCrop(224),
+            transforms.RandomHorizontalFlip(),
+            transforms.ToTensor()
+        ]),
+        transforms.Compose([
+            transforms.RandomRotation(15),
+            transforms.ColorJitter(brightness=0.3, contrast=0.3),
+            transforms.ToTensor()
+        ]),
+        transforms.Compose([
+            transforms.RandomAffine(degrees=20, scale=(0.8, 1.2)),
+            transforms.ToTensor()
+        ]),
+        transforms.Compose([
+            transforms.RandomGrayscale(p=0.2),
+            transforms.RandomHorizontalFlip(),
+            transforms.ToTensor()
+        ]),
+        transforms.Compose([
+            # transforms.RandomResizedCrop(224),
+            # transforms.RandomVerticalFlip(),
+            transforms.ToTensor()
+        ])
+    ]
+
+    args = parser.parse_args()
+    print(args)
+    imagenet_classes = ["tench", "goldfish", "great white shark", "tiger shark", "hammerhead shark", "electric ray",
+                        "stingray", "rooster", "hen", "ostrich", "brambling", "goldfinch", "house finch", "junco",
+                        "indigo bunting", "American robin", "bulbul", "jay", "magpie", "chickadee", "American dipper",
+                        "kite (bird of prey)", "bald eagle", "vulture", "great grey owl", "fire salamander",
+                        "smooth newt", "newt", "spotted salamander", "axolotl", "American bullfrog", "tree frog",
+                        "tailed frog", "loggerhead sea turtle", "leatherback sea turtle", "mud turtle", "terrapin",
+                        "box turtle", "banded gecko", "green iguana", "Carolina anole",
+                        "desert grassland whiptail lizard", "agama", "frilled-necked lizard", "alligator lizard",
+                        "Gila monster", "European green lizard", "chameleon", "Komodo dragon", "Nile crocodile",
+                        "American alligator", "triceratops", "worm snake", "ring-necked snake",
+                        "eastern hog-nosed snake", "smooth green snake", "kingsnake", "garter snake", "water snake",
+                        "vine snake", "night snake", "boa constrictor", "African rock python", "Indian cobra",
+                        "green mamba", "sea snake", "Saharan horned viper", "eastern diamondback rattlesnake",
+                        "sidewinder rattlesnake", "trilobite", "harvestman", "scorpion", "yellow garden spider",
+                        "barn spider", "European garden spider", "southern black widow", "tarantula", "wolf spider",
+                        "tick", "centipede", "black grouse", "ptarmigan", "ruffed grouse", "prairie grouse", "peafowl",
+                        "quail", "partridge", "african grey parrot", "macaw", "sulphur-crested cockatoo", "lorikeet",
+                        "coucal", "bee eater", "hornbill", "hummingbird", "jacamar", "toucan", "duck",
+                        "red-breasted merganser", "goose", "black swan", "tusker", "echidna", "platypus", "wallaby",
+                        "koala", "wombat", "jellyfish", "sea anemone", "brain coral", "flatworm", "nematode", "conch",
+                        "snail", "slug", "sea slug", "chiton", "chambered nautilus", "Dungeness crab", "rock crab",
+                        "fiddler crab", "red king crab", "American lobster", "spiny lobster", "crayfish", "hermit crab",
+                        "isopod", "white stork", "black stork", "spoonbill", "flamingo", "little blue heron",
+                        "great egret", "bittern bird", "crane bird", "limpkin", "common gallinule", "American coot",
+                        "bustard", "ruddy turnstone", "dunlin", "common redshank", "dowitcher", "oystercatcher",
+                        "pelican", "king penguin", "albatross", "grey whale", "killer whale", "dugong", "sea lion",
+                        "Chihuahua", "Japanese Chin", "Maltese", "Pekingese", "Shih Tzu", "King Charles Spaniel",
+                        "Papillon", "toy terrier", "Rhodesian Ridgeback", "Afghan Hound", "Basset Hound", "Beagle",
+                        "Bloodhound", "Bluetick Coonhound", "Black and Tan Coonhound", "Treeing Walker Coonhound",
+                        "English foxhound", "Redbone Coonhound", "borzoi", "Irish Wolfhound", "Italian Greyhound",
+                        "Whippet", "Ibizan Hound", "Norwegian Elkhound", "Otterhound", "Saluki", "Scottish Deerhound",
+                        "Weimaraner", "Staffordshire Bull Terrier", "American Staffordshire Terrier",
+                        "Bedlington Terrier", "Border Terrier", "Kerry Blue Terrier", "Irish Terrier",
+                        "Norfolk Terrier", "Norwich Terrier", "Yorkshire Terrier", "Wire Fox Terrier",
+                        "Lakeland Terrier", "Sealyham Terrier", "Airedale Terrier", "Cairn Terrier",
+                        "Australian Terrier", "Dandie Dinmont Terrier", "Boston Terrier", "Miniature Schnauzer",
+                        "Giant Schnauzer", "Standard Schnauzer", "Scottish Terrier", "Tibetan Terrier",
+                        "Australian Silky Terrier", "Soft-coated Wheaten Terrier", "West Highland White Terrier",
+                        "Lhasa Apso", "Flat-Coated Retriever", "Curly-coated Retriever", "Golden Retriever",
+                        "Labrador Retriever", "Chesapeake Bay Retriever", "German Shorthaired Pointer", "Vizsla",
+                        "English Setter", "Irish Setter", "Gordon Setter", "Brittany dog", "Clumber Spaniel",
+                        "English Springer Spaniel", "Welsh Springer Spaniel", "Cocker Spaniel", "Sussex Spaniel",
+                        "Irish Water Spaniel", "Kuvasz", "Schipperke", "Groenendael dog", "Malinois", "Briard",
+                        "Australian Kelpie", "Komondor", "Old English Sheepdog", "Shetland Sheepdog", "collie",
+                        "Border Collie", "Bouvier des Flandres dog", "Rottweiler", "German Shepherd Dog", "Dobermann",
+                        "Miniature Pinscher", "Greater Swiss Mountain Dog", "Bernese Mountain Dog",
+                        "Appenzeller Sennenhund", "Entlebucher Sennenhund", "Boxer", "Bullmastiff", "Tibetan Mastiff",
+                        "French Bulldog", "Great Dane", "St. Bernard", "husky", "Alaskan Malamute", "Siberian Husky",
+                        "Dalmatian", "Affenpinscher", "Basenji", "pug", "Leonberger", "Newfoundland dog",
+                        "Great Pyrenees dog", "Samoyed", "Pomeranian", "Chow Chow", "Keeshond", "brussels griffon",
+                        "Pembroke Welsh Corgi", "Cardigan Welsh Corgi", "Toy Poodle", "Miniature Poodle",
+                        "Standard Poodle", "Mexican hairless dog (xoloitzcuintli)", "grey wolf", "Alaskan tundra wolf",
+                        "red wolf or maned wolf", "coyote", "dingo", "dhole", "African wild dog", "hyena", "red fox",
+                        "kit fox", "Arctic fox", "grey fox", "tabby cat", "tiger cat", "Persian cat", "Siamese cat",
+                        "Egyptian Mau", "cougar", "lynx", "leopard", "snow leopard", "jaguar", "lion", "tiger",
+                        "cheetah", "brown bear", "American black bear", "polar bear", "sloth bear", "mongoose",
+                        "meerkat", "tiger beetle", "ladybug", "ground beetle", "longhorn beetle", "leaf beetle",
+                        "dung beetle", "rhinoceros beetle", "weevil", "fly", "bee", "ant", "grasshopper",
+                        "cricket insect", "stick insect", "cockroach", "praying mantis", "cicada", "leafhopper",
+                        "lacewing", "dragonfly", "damselfly", "red admiral butterfly", "ringlet butterfly",
+                        "monarch butterfly", "small white butterfly", "sulphur butterfly", "gossamer-winged butterfly",
+                        "starfish", "sea urchin", "sea cucumber", "cottontail rabbit", "hare", "Angora rabbit",
+                        "hamster", "porcupine", "fox squirrel", "marmot", "beaver", "guinea pig", "common sorrel horse",
+                        "zebra", "pig", "wild boar", "warthog", "hippopotamus", "ox", "water buffalo", "bison",
+                        "ram (adult male sheep)", "bighorn sheep", "Alpine ibex", "hartebeest", "impala (antelope)",
+                        "gazelle", "arabian camel", "llama", "weasel", "mink", "European polecat",
+                        "black-footed ferret", "otter", "skunk", "badger", "armadillo", "three-toed sloth", "orangutan",
+                        "gorilla", "chimpanzee", "gibbon", "siamang", "guenon", "patas monkey", "baboon", "macaque",
+                        "langur", "black-and-white colobus", "proboscis monkey", "marmoset", "white-headed capuchin",
+                        "howler monkey", "titi monkey", "Geoffroy's spider monkey", "common squirrel monkey",
+                        "ring-tailed lemur", "indri", "Asian elephant", "African bush elephant", "red panda",
+                        "giant panda", "snoek fish", "eel", "silver salmon", "rock beauty fish", "clownfish",
+                        "sturgeon", "gar fish", "lionfish", "pufferfish", "abacus", "abaya", "academic gown",
+                        "accordion", "acoustic guitar", "aircraft carrier", "airliner", "airship", "altar", "ambulance",
+                        "amphibious vehicle", "analog clock", "apiary", "apron", "trash can", "assault rifle",
+                        "backpack", "bakery", "balance beam", "balloon", "ballpoint pen", "Band-Aid", "banjo",
+                        "baluster / handrail", "barbell", "barber chair", "barbershop", "barn", "barometer", "barrel",
+                        "wheelbarrow", "baseball", "basketball", "bassinet", "bassoon", "swimming cap", "bath towel",
+                        "bathtub", "station wagon", "lighthouse", "beaker", "military hat (bearskin or shako)",
+                        "beer bottle", "beer glass", "bell tower", "baby bib", "tandem bicycle", "bikini",
+                        "ring binder", "binoculars", "birdhouse", "boathouse", "bobsleigh", "bolo tie", "poke bonnet",
+                        "bookcase", "bookstore", "bottle cap", "hunting bow", "bow tie", "brass memorial plaque", "bra",
+                        "breakwater", "breastplate", "broom", "bucket", "buckle", "bulletproof vest",
+                        "high-speed train", "butcher shop", "taxicab", "cauldron", "candle", "cannon", "canoe",
+                        "can opener", "cardigan", "car mirror", "carousel", "tool kit", "cardboard box / carton",
+                        "car wheel", "automated teller machine", "cassette", "cassette player", "castle", "catamaran",
+                        "CD player", "cello", "mobile phone", "chain", "chain-link fence", "chain mail", "chainsaw",
+                        "storage chest", "chiffonier", "bell or wind chime", "china cabinet", "Christmas stocking",
+                        "church", "movie theater", "cleaver", "cliff dwelling", "cloak", "clogs", "cocktail shaker",
+                        "coffee mug", "coffeemaker", "spiral or coil", "combination lock", "computer keyboard",
+                        "candy store", "container ship", "convertible", "corkscrew", "cornet", "cowboy boot",
+                        "cowboy hat", "cradle", "construction crane", "crash helmet", "crate", "infant bed",
+                        "Crock Pot", "croquet ball", "crutch", "cuirass", "dam", "desk", "desktop computer",
+                        "rotary dial telephone", "diaper", "digital clock", "digital watch", "dining table",
+                        "dishcloth", "dishwasher", "disc brake", "dock", "dog sled", "dome", "doormat", "drilling rig",
+                        "drum", "drumstick", "dumbbell", "Dutch oven", "electric fan", "electric guitar",
+                        "electric locomotive", "entertainment center", "envelope", "espresso machine", "face powder",
+                        "feather boa", "filing cabinet", "fireboat", "fire truck", "fire screen", "flagpole", "flute",
+                        "folding chair", "football helmet", "forklift", "fountain", "fountain pen", "four-poster bed",
+                        "freight car", "French horn", "frying pan", "fur coat", "garbage truck",
+                        "gas mask or respirator", "gas pump", "goblet", "go-kart", "golf ball", "golf cart", "gondola",
+                        "gong", "gown", "grand piano", "greenhouse", "radiator grille", "grocery store", "guillotine",
+                        "hair clip", "hair spray", "half-track", "hammer", "hamper", "hair dryer", "hand-held computer",
+                        "handkerchief", "hard disk drive", "harmonica", "harp", "combine harvester", "hatchet",
+                        "holster", "home theater", "honeycomb", "hook", "hoop skirt", "gymnastic horizontal bar",
+                        "horse-drawn vehicle", "hourglass", "iPod", "clothes iron", "carved pumpkin", "jeans", "jeep",
+                        "T-shirt", "jigsaw puzzle", "rickshaw", "joystick", "kimono", "knee pad", "knot", "lab coat",
+                        "ladle", "lampshade", "laptop computer", "lawn mower", "lens cap", "letter opener", "library",
+                        "lifeboat", "lighter", "limousine", "ocean liner", "lipstick", "slip-on shoe", "lotion",
+                        "music speaker", "loupe magnifying glass", "sawmill", "magnetic compass", "messenger bag",
+                        "mailbox", "tights", "one-piece bathing suit", "manhole cover", "maraca", "marimba", "mask",
+                        "matchstick", "maypole", "maze", "measuring cup", "medicine cabinet", "megalith", "microphone",
+                        "microwave oven", "military uniform", "milk can", "minibus", "miniskirt", "minivan", "missile",
+                        "mitten", "mixing bowl", "mobile home", "ford model t", "modem", "monastery", "monitor",
+                        "moped", "mortar and pestle", "graduation cap", "mosque", "mosquito net", "vespa",
+                        "mountain bike", "tent", "computer mouse", "mousetrap", "moving van", "muzzle", "metal nail",
+                        "neck brace", "necklace", "baby pacifier", "notebook computer", "obelisk", "oboe", "ocarina",
+                        "odometer", "oil filter", "pipe organ", "oscilloscope", "overskirt", "bullock cart",
+                        "oxygen mask", "product packet / packaging", "paddle", "paddle wheel", "padlock", "paintbrush",
+                        "pajamas", "palace", "pan flute", "paper towel", "parachute", "parallel bars", "park bench",
+                        "parking meter", "railroad car", "patio", "payphone", "pedestal", "pencil case",
+                        "pencil sharpener", "perfume", "Petri dish", "photocopier", "plectrum", "Pickelhaube",
+                        "picket fence", "pickup truck", "pier", "piggy bank", "pill bottle", "pillow", "ping-pong ball",
+                        "pinwheel", "pirate ship", "drink pitcher", "block plane", "planetarium", "plastic bag",
+                        "plate rack", "farm plow", "plunger", "Polaroid camera", "pole", "police van", "poncho",
+                        "pool table", "soda bottle", "plant pot", "potter's wheel", "power drill", "prayer rug",
+                        "printer", "prison", "missile", "projector", "hockey puck", "punching bag", "purse", "quill",
+                        "quilt", "race car", "racket", "radiator", "radio", "radio telescope", "rain barrel",
+                        "recreational vehicle", "fishing casting reel", "reflex camera", "refrigerator",
+                        "remote control", "restaurant", "revolver", "rifle", "rocking chair", "rotisserie", "eraser",
+                        "rugby ball", "ruler measuring stick", "sneaker", "safe", "safety pin", "salt shaker", "sandal",
+                        "sarong", "saxophone", "scabbard", "weighing scale", "school bus", "schooner", "scoreboard",
+                        "CRT monitor", "screw", "screwdriver", "seat belt", "sewing machine", "shield", "shoe store",
+                        "shoji screen / room divider", "shopping basket", "shopping cart", "shovel", "shower cap",
+                        "shower curtain", "ski", "balaclava ski mask", "sleeping bag", "slide rule", "sliding door",
+                        "slot machine", "snorkel", "snowmobile", "snowplow", "soap dispenser", "soccer ball", "sock",
+                        "solar thermal collector", "sombrero", "soup bowl", "keyboard space bar", "space heater",
+                        "space shuttle", "spatula", "motorboat", "spider web", "spindle", "sports car", "spotlight",
+                        "stage", "steam locomotive", "through arch bridge", "steel drum", "stethoscope", "scarf",
+                        "stone wall", "stopwatch", "stove", "strainer", "tram", "stretcher", "couch", "stupa",
+                        "submarine", "suit", "sundial", "sunglasses", "sunglasses", "sunscreen", "suspension bridge",
+                        "mop", "sweatshirt", "swim trunks / shorts", "swing", "electrical switch", "syringe",
+                        "table lamp", "tank", "tape player", "teapot", "teddy bear", "television", "tennis ball",
+                        "thatched roof", "front curtain", "thimble", "threshing machine", "throne", "tile roof",
+                        "toaster", "tobacco shop", "toilet seat", "torch", "totem pole", "tow truck", "toy store",
+                        "tractor", "semi-trailer truck", "tray", "trench coat", "tricycle", "trimaran", "tripod",
+                        "triumphal arch", "trolleybus", "trombone", "hot tub", "turnstile", "typewriter keyboard",
+                        "umbrella", "unicycle", "upright piano", "vacuum cleaner", "vase", "vaulted or arched ceiling",
+                        "velvet fabric", "vending machine", "vestment", "viaduct", "violin", "volleyball",
+                        "waffle iron", "wall clock", "wallet", "wardrobe", "military aircraft", "sink",
+                        "washing machine", "water bottle", "water jug", "water tower", "whiskey jug", "whistle",
+                        "hair wig", "window screen", "window shade", "Windsor tie", "wine bottle", "airplane wing",
+                        "wok", "wooden spoon", "wool", "split-rail fence", "shipwreck", "sailboat", "yurt", "website",
+                        "comic book", "crossword", "traffic or street sign", "traffic light", "dust jacket", "menu",
+                        "plate", "guacamole", "consomme", "hot pot", "trifle", "ice cream", "popsicle", "baguette",
+                        "bagel", "pretzel", "cheeseburger", "hot dog", "mashed potatoes", "cabbage", "broccoli",
+                        "cauliflower", "zucchini", "spaghetti squash", "acorn squash", "butternut squash", "cucumber",
+                        "artichoke", "bell pepper", "cardoon", "mushroom", "Granny Smith apple", "strawberry", "orange",
+                        "lemon", "fig", "pineapple", "banana", "jackfruit", "cherimoya (custard apple)", "pomegranate",
+                        "hay", "carbonara", "chocolate syrup", "dough", "meatloaf", "pizza", "pot pie", "burrito",
+                        "red wine", "espresso", "tea cup", "eggnog", "mountain", "bubble", "cliff", "coral reef",
+                        "geyser", "lakeshore", "promontory", "sandbar", "beach", "valley", "volcano", "baseball player",
+                        "bridegroom", "scuba diver", "rapeseed", "daisy", "yellow lady's slipper", "corn", "acorn",
+                        "rose hip", "horse chestnut seed", "coral fungus", "agaric", "gyromitra", "stinkhorn mushroom",
+                        "earth star fungus", "hen of the woods mushroom", "bolete", "corn cob", "toilet paper"]
+
+    # cifar100_classes = [name for name in os.listdir(args.data_path)
+    #                 if os.path.isdir(os.path.join(args.data_path, name))]
+
+    imagenet_single_template = [
+        'a photo of a {}.',
+    ]
+
+    imagenet_7_templates = [
+        'itap of a {}.',
+        'a origami {}.',
+        'a bad photo of the {}.',
+        'a photo of the large {}.',
+        'a {} in a video game.',
+        'art of the {}.',
+        'a photo of the small {}.',
+    ]
+
+
+    print('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    print('load data')
+    valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+
+    # aug 1
+    with torch.no_grad():
+        image_feat_0 = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            # target = target.cuda()
+            augmented_images_0 = torch.stack([
+                augmentations[0](to_pil_image(image.cpu())) for image in images
+            ]).cuda()
+            image_features_0 = model.encode_image(augmented_images_0)
+            image_feat_0.append(F.normalize(image_features_0, dim=1))
+            # image_label.append(target)
+    image_feat_0 = torch.cat(image_feat_0, dim=0)
+
+    # aug2
+    with torch.no_grad():
+        image_feat_1 = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            # target = target.cuda()
+            augmented_images_1 = torch.stack([
+                augmentations[1](to_pil_image(image.cpu())) for image in images
+            ]).cuda()
+            image_features_1 = model.encode_image(augmented_images_1)
+            image_feat_1.append(F.normalize(image_features_1, dim=1))
+            # image_label.append(target)
+    image_feat_1 = torch.cat(image_feat_1, dim=0)
+
+    # aug3
+    with torch.no_grad():
+        image_feat_2 = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            # target = target.cuda()
+            augmented_images_2 = torch.stack([
+                augmentations[2](to_pil_image(image.cpu())) for image in images
+            ]).cuda()
+            image_features_2 = model.encode_image(augmented_images_2)
+            image_feat_2.append(F.normalize(image_features_2, dim=1))
+            # image_label.append(target)
+    image_feat_2 = torch.cat(image_feat_2, dim=0)
+
+    # aug4
+    with torch.no_grad():
+        image_feat_3 = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            # target = target.cuda()
+            augmented_images_3 = torch.stack([
+                augmentations[3](to_pil_image(image.cpu())) for image in images
+            ]).cuda()
+            image_features_3 = model.encode_image(augmented_images_3)
+            image_feat_3.append(F.normalize(image_features_3, dim=1))
+            # image_label.append(target)
+    image_feat_3 = torch.cat(image_feat_3, dim=0)
+
+    # aug5
+    with torch.no_grad():
+        image_feat_4 = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            # target = target.cuda()
+            augmented_images_4 = torch.stack([
+                augmentations[4](to_pil_image(image.cpu())) for image in images
+            ]).cuda()
+            image_features_4 = model.encode_image(augmented_images_4)
+            image_feat_4.append(F.normalize(image_features_4, dim=1))
+            # image_label.append(target)
+    image_feat_4 = torch.cat(image_feat_4, dim=0)
+
+    image_feat = torch.mean(torch.stack([image_feat_4], dim=0), dim=0)
+
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    print('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, imagenet_classes, imagenet_7_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    print(f'accuracy with text proxy: {top1:.2f}')
+
+    print('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+            beta = args.beta * math.sqrt((i + 1) / n)
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+            rho -= rlr * (tlabel - args.alpha / num_class)
+            rho[rho < 0] = 0
+            text_label[i, :] = tlabel
+            vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            grad = torch.outer(x, vision_label - tlabel)
+            w -= (lr / args.tau_i) * grad
+            w = F.normalize(w, dim=0)
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+
+        # MAPLS - EM Algorithm
+        pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+        qy = mapls(combo_label, pz = pz, qy_mode = "soft", max_iter = 100, lam = lam)
+
+        w = np.array(qy) / np.array(pz)
+        if combo_label.is_cuda:
+            combo_label_cpu = combo_label.cpu()
+        qy_probs = lsc(combo_label_cpu, 1.0/w)
+        qy_probs = torch.from_numpy(qy_probs)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+    print('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    print('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    print('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    lams = [0.6]
+    for lam in lams:
+        main(lam)
+

OnZeta/main_online_imagenet_mapls_inloop.py

@@ -0,0 +1,347 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='/home/li325/space_mlai/pengxiao_space/dataset/ImageNet/', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='RN50',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=256, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.8, type=float)
+parser.add_argument('--repeat', default=1, type=int)
+
+# parser.add_argument('--lam', default=0.6, type=float)
+
+
+def main(lam):
+
+    args = parser.parse_args()
+    print(args)
+    print("lambda = ", lam)
+    imagenet_classes = ["tench", "goldfish", "great white shark", "tiger shark", "hammerhead shark", "electric ray",
+                        "stingray", "rooster", "hen", "ostrich", "brambling", "goldfinch", "house finch", "junco",
+                        "indigo bunting", "American robin", "bulbul", "jay", "magpie", "chickadee", "American dipper",
+                        "kite (bird of prey)", "bald eagle", "vulture", "great grey owl", "fire salamander",
+                        "smooth newt", "newt", "spotted salamander", "axolotl", "American bullfrog", "tree frog",
+                        "tailed frog", "loggerhead sea turtle", "leatherback sea turtle", "mud turtle", "terrapin",
+                        "box turtle", "banded gecko", "green iguana", "Carolina anole",
+                        "desert grassland whiptail lizard", "agama", "frilled-necked lizard", "alligator lizard",
+                        "Gila monster", "European green lizard", "chameleon", "Komodo dragon", "Nile crocodile",
+                        "American alligator", "triceratops", "worm snake", "ring-necked snake",
+                        "eastern hog-nosed snake", "smooth green snake", "kingsnake", "garter snake", "water snake",
+                        "vine snake", "night snake", "boa constrictor", "African rock python", "Indian cobra",
+                        "green mamba", "sea snake", "Saharan horned viper", "eastern diamondback rattlesnake",
+                        "sidewinder rattlesnake", "trilobite", "harvestman", "scorpion", "yellow garden spider",
+                        "barn spider", "European garden spider", "southern black widow", "tarantula", "wolf spider",
+                        "tick", "centipede", "black grouse", "ptarmigan", "ruffed grouse", "prairie grouse", "peafowl",
+                        "quail", "partridge", "african grey parrot", "macaw", "sulphur-crested cockatoo", "lorikeet",
+                        "coucal", "bee eater", "hornbill", "hummingbird", "jacamar", "toucan", "duck",
+                        "red-breasted merganser", "goose", "black swan", "tusker", "echidna", "platypus", "wallaby",
+                        "koala", "wombat", "jellyfish", "sea anemone", "brain coral", "flatworm", "nematode", "conch",
+                        "snail", "slug", "sea slug", "chiton", "chambered nautilus", "Dungeness crab", "rock crab",
+                        "fiddler crab", "red king crab", "American lobster", "spiny lobster", "crayfish", "hermit crab",
+                        "isopod", "white stork", "black stork", "spoonbill", "flamingo", "little blue heron",
+                        "great egret", "bittern bird", "crane bird", "limpkin", "common gallinule", "American coot",
+                        "bustard", "ruddy turnstone", "dunlin", "common redshank", "dowitcher", "oystercatcher",
+                        "pelican", "king penguin", "albatross", "grey whale", "killer whale", "dugong", "sea lion",
+                        "Chihuahua", "Japanese Chin", "Maltese", "Pekingese", "Shih Tzu", "King Charles Spaniel",
+                        "Papillon", "toy terrier", "Rhodesian Ridgeback", "Afghan Hound", "Basset Hound", "Beagle",
+                        "Bloodhound", "Bluetick Coonhound", "Black and Tan Coonhound", "Treeing Walker Coonhound",
+                        "English foxhound", "Redbone Coonhound", "borzoi", "Irish Wolfhound", "Italian Greyhound",
+                        "Whippet", "Ibizan Hound", "Norwegian Elkhound", "Otterhound", "Saluki", "Scottish Deerhound",
+                        "Weimaraner", "Staffordshire Bull Terrier", "American Staffordshire Terrier",
+                        "Bedlington Terrier", "Border Terrier", "Kerry Blue Terrier", "Irish Terrier",
+                        "Norfolk Terrier", "Norwich Terrier", "Yorkshire Terrier", "Wire Fox Terrier",
+                        "Lakeland Terrier", "Sealyham Terrier", "Airedale Terrier", "Cairn Terrier",
+                        "Australian Terrier", "Dandie Dinmont Terrier", "Boston Terrier", "Miniature Schnauzer",
+                        "Giant Schnauzer", "Standard Schnauzer", "Scottish Terrier", "Tibetan Terrier",
+                        "Australian Silky Terrier", "Soft-coated Wheaten Terrier", "West Highland White Terrier",
+                        "Lhasa Apso", "Flat-Coated Retriever", "Curly-coated Retriever", "Golden Retriever",
+                        "Labrador Retriever", "Chesapeake Bay Retriever", "German Shorthaired Pointer", "Vizsla",
+                        "English Setter", "Irish Setter", "Gordon Setter", "Brittany dog", "Clumber Spaniel",
+                        "English Springer Spaniel", "Welsh Springer Spaniel", "Cocker Spaniel", "Sussex Spaniel",
+                        "Irish Water Spaniel", "Kuvasz", "Schipperke", "Groenendael dog", "Malinois", "Briard",
+                        "Australian Kelpie", "Komondor", "Old English Sheepdog", "Shetland Sheepdog", "collie",
+                        "Border Collie", "Bouvier des Flandres dog", "Rottweiler", "German Shepherd Dog", "Dobermann",
+                        "Miniature Pinscher", "Greater Swiss Mountain Dog", "Bernese Mountain Dog",
+                        "Appenzeller Sennenhund", "Entlebucher Sennenhund", "Boxer", "Bullmastiff", "Tibetan Mastiff",
+                        "French Bulldog", "Great Dane", "St. Bernard", "husky", "Alaskan Malamute", "Siberian Husky",
+                        "Dalmatian", "Affenpinscher", "Basenji", "pug", "Leonberger", "Newfoundland dog",
+                        "Great Pyrenees dog", "Samoyed", "Pomeranian", "Chow Chow", "Keeshond", "brussels griffon",
+                        "Pembroke Welsh Corgi", "Cardigan Welsh Corgi", "Toy Poodle", "Miniature Poodle",
+                        "Standard Poodle", "Mexican hairless dog (xoloitzcuintli)", "grey wolf", "Alaskan tundra wolf",
+                        "red wolf or maned wolf", "coyote", "dingo", "dhole", "African wild dog", "hyena", "red fox",
+                        "kit fox", "Arctic fox", "grey fox", "tabby cat", "tiger cat", "Persian cat", "Siamese cat",
+                        "Egyptian Mau", "cougar", "lynx", "leopard", "snow leopard", "jaguar", "lion", "tiger",
+                        "cheetah", "brown bear", "American black bear", "polar bear", "sloth bear", "mongoose",
+                        "meerkat", "tiger beetle", "ladybug", "ground beetle", "longhorn beetle", "leaf beetle",
+                        "dung beetle", "rhinoceros beetle", "weevil", "fly", "bee", "ant", "grasshopper",
+                        "cricket insect", "stick insect", "cockroach", "praying mantis", "cicada", "leafhopper",
+                        "lacewing", "dragonfly", "damselfly", "red admiral butterfly", "ringlet butterfly",
+                        "monarch butterfly", "small white butterfly", "sulphur butterfly", "gossamer-winged butterfly",
+                        "starfish", "sea urchin", "sea cucumber", "cottontail rabbit", "hare", "Angora rabbit",
+                        "hamster", "porcupine", "fox squirrel", "marmot", "beaver", "guinea pig", "common sorrel horse",
+                        "zebra", "pig", "wild boar", "warthog", "hippopotamus", "ox", "water buffalo", "bison",
+                        "ram (adult male sheep)", "bighorn sheep", "Alpine ibex", "hartebeest", "impala (antelope)",
+                        "gazelle", "arabian camel", "llama", "weasel", "mink", "European polecat",
+                        "black-footed ferret", "otter", "skunk", "badger", "armadillo", "three-toed sloth", "orangutan",
+                        "gorilla", "chimpanzee", "gibbon", "siamang", "guenon", "patas monkey", "baboon", "macaque",
+                        "langur", "black-and-white colobus", "proboscis monkey", "marmoset", "white-headed capuchin",
+                        "howler monkey", "titi monkey", "Geoffroy's spider monkey", "common squirrel monkey",
+                        "ring-tailed lemur", "indri", "Asian elephant", "African bush elephant", "red panda",
+                        "giant panda", "snoek fish", "eel", "silver salmon", "rock beauty fish", "clownfish",
+                        "sturgeon", "gar fish", "lionfish", "pufferfish", "abacus", "abaya", "academic gown",
+                        "accordion", "acoustic guitar", "aircraft carrier", "airliner", "airship", "altar", "ambulance",
+                        "amphibious vehicle", "analog clock", "apiary", "apron", "trash can", "assault rifle",
+                        "backpack", "bakery", "balance beam", "balloon", "ballpoint pen", "Band-Aid", "banjo",
+                        "baluster / handrail", "barbell", "barber chair", "barbershop", "barn", "barometer", "barrel",
+                        "wheelbarrow", "baseball", "basketball", "bassinet", "bassoon", "swimming cap", "bath towel",
+                        "bathtub", "station wagon", "lighthouse", "beaker", "military hat (bearskin or shako)",
+                        "beer bottle", "beer glass", "bell tower", "baby bib", "tandem bicycle", "bikini",
+                        "ring binder", "binoculars", "birdhouse", "boathouse", "bobsleigh", "bolo tie", "poke bonnet",
+                        "bookcase", "bookstore", "bottle cap", "hunting bow", "bow tie", "brass memorial plaque", "bra",
+                        "breakwater", "breastplate", "broom", "bucket", "buckle", "bulletproof vest",
+                        "high-speed train", "butcher shop", "taxicab", "cauldron", "candle", "cannon", "canoe",
+                        "can opener", "cardigan", "car mirror", "carousel", "tool kit", "cardboard box / carton",
+                        "car wheel", "automated teller machine", "cassette", "cassette player", "castle", "catamaran",
+                        "CD player", "cello", "mobile phone", "chain", "chain-link fence", "chain mail", "chainsaw",
+                        "storage chest", "chiffonier", "bell or wind chime", "china cabinet", "Christmas stocking",
+                        "church", "movie theater", "cleaver", "cliff dwelling", "cloak", "clogs", "cocktail shaker",
+                        "coffee mug", "coffeemaker", "spiral or coil", "combination lock", "computer keyboard",
+                        "candy store", "container ship", "convertible", "corkscrew", "cornet", "cowboy boot",
+                        "cowboy hat", "cradle", "construction crane", "crash helmet", "crate", "infant bed",
+                        "Crock Pot", "croquet ball", "crutch", "cuirass", "dam", "desk", "desktop computer",
+                        "rotary dial telephone", "diaper", "digital clock", "digital watch", "dining table",
+                        "dishcloth", "dishwasher", "disc brake", "dock", "dog sled", "dome", "doormat", "drilling rig",
+                        "drum", "drumstick", "dumbbell", "Dutch oven", "electric fan", "electric guitar",
+                        "electric locomotive", "entertainment center", "envelope", "espresso machine", "face powder",
+                        "feather boa", "filing cabinet", "fireboat", "fire truck", "fire screen", "flagpole", "flute",
+                        "folding chair", "football helmet", "forklift", "fountain", "fountain pen", "four-poster bed",
+                        "freight car", "French horn", "frying pan", "fur coat", "garbage truck",
+                        "gas mask or respirator", "gas pump", "goblet", "go-kart", "golf ball", "golf cart", "gondola",
+                        "gong", "gown", "grand piano", "greenhouse", "radiator grille", "grocery store", "guillotine",
+                        "hair clip", "hair spray", "half-track", "hammer", "hamper", "hair dryer", "hand-held computer",
+                        "handkerchief", "hard disk drive", "harmonica", "harp", "combine harvester", "hatchet",
+                        "holster", "home theater", "honeycomb", "hook", "hoop skirt", "gymnastic horizontal bar",
+                        "horse-drawn vehicle", "hourglass", "iPod", "clothes iron", "carved pumpkin", "jeans", "jeep",
+                        "T-shirt", "jigsaw puzzle", "rickshaw", "joystick", "kimono", "knee pad", "knot", "lab coat",
+                        "ladle", "lampshade", "laptop computer", "lawn mower", "lens cap", "letter opener", "library",
+                        "lifeboat", "lighter", "limousine", "ocean liner", "lipstick", "slip-on shoe", "lotion",
+                        "music speaker", "loupe magnifying glass", "sawmill", "magnetic compass", "messenger bag",
+                        "mailbox", "tights", "one-piece bathing suit", "manhole cover", "maraca", "marimba", "mask",
+                        "matchstick", "maypole", "maze", "measuring cup", "medicine cabinet", "megalith", "microphone",
+                        "microwave oven", "military uniform", "milk can", "minibus", "miniskirt", "minivan", "missile",
+                        "mitten", "mixing bowl", "mobile home", "ford model t", "modem", "monastery", "monitor",
+                        "moped", "mortar and pestle", "graduation cap", "mosque", "mosquito net", "vespa",
+                        "mountain bike", "tent", "computer mouse", "mousetrap", "moving van", "muzzle", "metal nail",
+                        "neck brace", "necklace", "baby pacifier", "notebook computer", "obelisk", "oboe", "ocarina",
+                        "odometer", "oil filter", "pipe organ", "oscilloscope", "overskirt", "bullock cart",
+                        "oxygen mask", "product packet / packaging", "paddle", "paddle wheel", "padlock", "paintbrush",
+                        "pajamas", "palace", "pan flute", "paper towel", "parachute", "parallel bars", "park bench",
+                        "parking meter", "railroad car", "patio", "payphone", "pedestal", "pencil case",
+                        "pencil sharpener", "perfume", "Petri dish", "photocopier", "plectrum", "Pickelhaube",
+                        "picket fence", "pickup truck", "pier", "piggy bank", "pill bottle", "pillow", "ping-pong ball",
+                        "pinwheel", "pirate ship", "drink pitcher", "block plane", "planetarium", "plastic bag",
+                        "plate rack", "farm plow", "plunger", "Polaroid camera", "pole", "police van", "poncho",
+                        "pool table", "soda bottle", "plant pot", "potter's wheel", "power drill", "prayer rug",
+                        "printer", "prison", "missile", "projector", "hockey puck", "punching bag", "purse", "quill",
+                        "quilt", "race car", "racket", "radiator", "radio", "radio telescope", "rain barrel",
+                        "recreational vehicle", "fishing casting reel", "reflex camera", "refrigerator",
+                        "remote control", "restaurant", "revolver", "rifle", "rocking chair", "rotisserie", "eraser",
+                        "rugby ball", "ruler measuring stick", "sneaker", "safe", "safety pin", "salt shaker", "sandal",
+                        "sarong", "saxophone", "scabbard", "weighing scale", "school bus", "schooner", "scoreboard",
+                        "CRT monitor", "screw", "screwdriver", "seat belt", "sewing machine", "shield", "shoe store",
+                        "shoji screen / room divider", "shopping basket", "shopping cart", "shovel", "shower cap",
+                        "shower curtain", "ski", "balaclava ski mask", "sleeping bag", "slide rule", "sliding door",
+                        "slot machine", "snorkel", "snowmobile", "snowplow", "soap dispenser", "soccer ball", "sock",
+                        "solar thermal collector", "sombrero", "soup bowl", "keyboard space bar", "space heater",
+                        "space shuttle", "spatula", "motorboat", "spider web", "spindle", "sports car", "spotlight",
+                        "stage", "steam locomotive", "through arch bridge", "steel drum", "stethoscope", "scarf",
+                        "stone wall", "stopwatch", "stove", "strainer", "tram", "stretcher", "couch", "stupa",
+                        "submarine", "suit", "sundial", "sunglasses", "sunglasses", "sunscreen", "suspension bridge",
+                        "mop", "sweatshirt", "swim trunks / shorts", "swing", "electrical switch", "syringe",
+                        "table lamp", "tank", "tape player", "teapot", "teddy bear", "television", "tennis ball",
+                        "thatched roof", "front curtain", "thimble", "threshing machine", "throne", "tile roof",
+                        "toaster", "tobacco shop", "toilet seat", "torch", "totem pole", "tow truck", "toy store",
+                        "tractor", "semi-trailer truck", "tray", "trench coat", "tricycle", "trimaran", "tripod",
+                        "triumphal arch", "trolleybus", "trombone", "hot tub", "turnstile", "typewriter keyboard",
+                        "umbrella", "unicycle", "upright piano", "vacuum cleaner", "vase", "vaulted or arched ceiling",
+                        "velvet fabric", "vending machine", "vestment", "viaduct", "violin", "volleyball",
+                        "waffle iron", "wall clock", "wallet", "wardrobe", "military aircraft", "sink",
+                        "washing machine", "water bottle", "water jug", "water tower", "whiskey jug", "whistle",
+                        "hair wig", "window screen", "window shade", "Windsor tie", "wine bottle", "airplane wing",
+                        "wok", "wooden spoon", "wool", "split-rail fence", "shipwreck", "sailboat", "yurt", "website",
+                        "comic book", "crossword", "traffic or street sign", "traffic light", "dust jacket", "menu",
+                        "plate", "guacamole", "consomme", "hot pot", "trifle", "ice cream", "popsicle", "baguette",
+                        "bagel", "pretzel", "cheeseburger", "hot dog", "mashed potatoes", "cabbage", "broccoli",
+                        "cauliflower", "zucchini", "spaghetti squash", "acorn squash", "butternut squash", "cucumber",
+                        "artichoke", "bell pepper", "cardoon", "mushroom", "Granny Smith apple", "strawberry", "orange",
+                        "lemon", "fig", "pineapple", "banana", "jackfruit", "cherimoya (custard apple)", "pomegranate",
+                        "hay", "carbonara", "chocolate syrup", "dough", "meatloaf", "pizza", "pot pie", "burrito",
+                        "red wine", "espresso", "tea cup", "eggnog", "mountain", "bubble", "cliff", "coral reef",
+                        "geyser", "lakeshore", "promontory", "sandbar", "beach", "valley", "volcano", "baseball player",
+                        "bridegroom", "scuba diver", "rapeseed", "daisy", "yellow lady's slipper", "corn", "acorn",
+                        "rose hip", "horse chestnut seed", "coral fungus", "agaric", "gyromitra", "stinkhorn mushroom",
+                        "earth star fungus", "hen of the woods mushroom", "bolete", "corn cob", "toilet paper"]
+
+    # cifar100_classes = [name for name in os.listdir(args.data_path)
+    #                 if os.path.isdir(os.path.join(args.data_path, name))]
+
+    imagenet_single_template = [
+        'a photo of a {}.',
+    ]
+
+    imagenet_7_templates = [
+        'itap of a {}.',
+        'a origami {}.',
+        'a bad photo of the {}.',
+        'a photo of the large {}.',
+        'a {} in a video game.',
+        'art of the {}.',
+        'a photo of the small {}.',
+    ]
+
+    print('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    print('load data')
+    valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    print('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, imagenet_classes, imagenet_7_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    print(f'accuracy with text proxy: {top1:.2f}')
+
+    print('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+            beta = args.beta * math.sqrt((i + 1) / n)
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+            rho -= rlr * (tlabel - args.alpha / num_class)
+            rho[rho < 0] = 0
+            text_label[i, :] = tlabel
+            vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            # MAPLS - EM Algorithm
+            pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+            qy = mapls(combo_label, pz=pz, qy_mode="soft", max_iter=50, lam=lam)
+
+            w_mapls = np.array(qy) / np.array(pz)
+            if combo_label.is_cuda:
+                combo_label = combo_label.cpu()
+            qy_probs = lsc(combo_label, 1.0 / w_mapls)
+            # MAPLS - EM Algorithm
+            grad = torch.outer(x, vision_label - tlabel)
+            # v = beta_momentum * v + (1 - beta_momentum) * grad
+            w -= (lr / args.tau_i) * grad
+            w = F.normalize(w, dim=0)  # if normalization is desired
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+        # # MAPLS - EM Algorithm
+        # pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+        # qy = mapls(combo_label, pz = pz, qy_mode = "soft", max_iter = 100, lam = lam)  # FIXME why return nan
+        #
+        # w = np.array(qy) / np.array(pz)
+        # if combo_label.is_cuda:
+        #     combo_label_cpu = combo_label.cpu()
+        # qy_probs = lsc(combo_label_cpu, 1.0/w)
+        # # MAPLS - EM Algorithm
+        qy_probs = torch.from_numpy(qy_probs)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+    print('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    print('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    print('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    # lams = [0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1]
+    lams = [0.7]
+    for lam in lams:
+        main(lam)
+

OnZeta/main_online_imagenet_mapls_lame.py

@@ -0,0 +1,366 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+from lame.lame import LAME
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='/home/lt_test/projects/datasets/ImageNet/', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='RN50',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=256, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.8, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+
+
+def main(lam):
+
+    args = parser.parse_args()
+    args.beta = lam
+    print(args)
+    imagenet_classes = ["tench", "goldfish", "great white shark", "tiger shark", "hammerhead shark", "electric ray",
+                        "stingray", "rooster", "hen", "ostrich", "brambling", "goldfinch", "house finch", "junco",
+                        "indigo bunting", "American robin", "bulbul", "jay", "magpie", "chickadee", "American dipper",
+                        "kite (bird of prey)", "bald eagle", "vulture", "great grey owl", "fire salamander",
+                        "smooth newt", "newt", "spotted salamander", "axolotl", "American bullfrog", "tree frog",
+                        "tailed frog", "loggerhead sea turtle", "leatherback sea turtle", "mud turtle", "terrapin",
+                        "box turtle", "banded gecko", "green iguana", "Carolina anole",
+                        "desert grassland whiptail lizard", "agama", "frilled-necked lizard", "alligator lizard",
+                        "Gila monster", "European green lizard", "chameleon", "Komodo dragon", "Nile crocodile",
+                        "American alligator", "triceratops", "worm snake", "ring-necked snake",
+                        "eastern hog-nosed snake", "smooth green snake", "kingsnake", "garter snake", "water snake",
+                        "vine snake", "night snake", "boa constrictor", "African rock python", "Indian cobra",
+                        "green mamba", "sea snake", "Saharan horned viper", "eastern diamondback rattlesnake",
+                        "sidewinder rattlesnake", "trilobite", "harvestman", "scorpion", "yellow garden spider",
+                        "barn spider", "European garden spider", "southern black widow", "tarantula", "wolf spider",
+                        "tick", "centipede", "black grouse", "ptarmigan", "ruffed grouse", "prairie grouse", "peafowl",
+                        "quail", "partridge", "african grey parrot", "macaw", "sulphur-crested cockatoo", "lorikeet",
+                        "coucal", "bee eater", "hornbill", "hummingbird", "jacamar", "toucan", "duck",
+                        "red-breasted merganser", "goose", "black swan", "tusker", "echidna", "platypus", "wallaby",
+                        "koala", "wombat", "jellyfish", "sea anemone", "brain coral", "flatworm", "nematode", "conch",
+                        "snail", "slug", "sea slug", "chiton", "chambered nautilus", "Dungeness crab", "rock crab",
+                        "fiddler crab", "red king crab", "American lobster", "spiny lobster", "crayfish", "hermit crab",
+                        "isopod", "white stork", "black stork", "spoonbill", "flamingo", "little blue heron",
+                        "great egret", "bittern bird", "crane bird", "limpkin", "common gallinule", "American coot",
+                        "bustard", "ruddy turnstone", "dunlin", "common redshank", "dowitcher", "oystercatcher",
+                        "pelican", "king penguin", "albatross", "grey whale", "killer whale", "dugong", "sea lion",
+                        "Chihuahua", "Japanese Chin", "Maltese", "Pekingese", "Shih Tzu", "King Charles Spaniel",
+                        "Papillon", "toy terrier", "Rhodesian Ridgeback", "Afghan Hound", "Basset Hound", "Beagle",
+                        "Bloodhound", "Bluetick Coonhound", "Black and Tan Coonhound", "Treeing Walker Coonhound",
+                        "English foxhound", "Redbone Coonhound", "borzoi", "Irish Wolfhound", "Italian Greyhound",
+                        "Whippet", "Ibizan Hound", "Norwegian Elkhound", "Otterhound", "Saluki", "Scottish Deerhound",
+                        "Weimaraner", "Staffordshire Bull Terrier", "American Staffordshire Terrier",
+                        "Bedlington Terrier", "Border Terrier", "Kerry Blue Terrier", "Irish Terrier",
+                        "Norfolk Terrier", "Norwich Terrier", "Yorkshire Terrier", "Wire Fox Terrier",
+                        "Lakeland Terrier", "Sealyham Terrier", "Airedale Terrier", "Cairn Terrier",
+                        "Australian Terrier", "Dandie Dinmont Terrier", "Boston Terrier", "Miniature Schnauzer",
+                        "Giant Schnauzer", "Standard Schnauzer", "Scottish Terrier", "Tibetan Terrier",
+                        "Australian Silky Terrier", "Soft-coated Wheaten Terrier", "West Highland White Terrier",
+                        "Lhasa Apso", "Flat-Coated Retriever", "Curly-coated Retriever", "Golden Retriever",
+                        "Labrador Retriever", "Chesapeake Bay Retriever", "German Shorthaired Pointer", "Vizsla",
+                        "English Setter", "Irish Setter", "Gordon Setter", "Brittany dog", "Clumber Spaniel",
+                        "English Springer Spaniel", "Welsh Springer Spaniel", "Cocker Spaniel", "Sussex Spaniel",
+                        "Irish Water Spaniel", "Kuvasz", "Schipperke", "Groenendael dog", "Malinois", "Briard",
+                        "Australian Kelpie", "Komondor", "Old English Sheepdog", "Shetland Sheepdog", "collie",
+                        "Border Collie", "Bouvier des Flandres dog", "Rottweiler", "German Shepherd Dog", "Dobermann",
+                        "Miniature Pinscher", "Greater Swiss Mountain Dog", "Bernese Mountain Dog",
+                        "Appenzeller Sennenhund", "Entlebucher Sennenhund", "Boxer", "Bullmastiff", "Tibetan Mastiff",
+                        "French Bulldog", "Great Dane", "St. Bernard", "husky", "Alaskan Malamute", "Siberian Husky",
+                        "Dalmatian", "Affenpinscher", "Basenji", "pug", "Leonberger", "Newfoundland dog",
+                        "Great Pyrenees dog", "Samoyed", "Pomeranian", "Chow Chow", "Keeshond", "brussels griffon",
+                        "Pembroke Welsh Corgi", "Cardigan Welsh Corgi", "Toy Poodle", "Miniature Poodle",
+                        "Standard Poodle", "Mexican hairless dog (xoloitzcuintli)", "grey wolf", "Alaskan tundra wolf",
+                        "red wolf or maned wolf", "coyote", "dingo", "dhole", "African wild dog", "hyena", "red fox",
+                        "kit fox", "Arctic fox", "grey fox", "tabby cat", "tiger cat", "Persian cat", "Siamese cat",
+                        "Egyptian Mau", "cougar", "lynx", "leopard", "snow leopard", "jaguar", "lion", "tiger",
+                        "cheetah", "brown bear", "American black bear", "polar bear", "sloth bear", "mongoose",
+                        "meerkat", "tiger beetle", "ladybug", "ground beetle", "longhorn beetle", "leaf beetle",
+                        "dung beetle", "rhinoceros beetle", "weevil", "fly", "bee", "ant", "grasshopper",
+                        "cricket insect", "stick insect", "cockroach", "praying mantis", "cicada", "leafhopper",
+                        "lacewing", "dragonfly", "damselfly", "red admiral butterfly", "ringlet butterfly",
+                        "monarch butterfly", "small white butterfly", "sulphur butterfly", "gossamer-winged butterfly",
+                        "starfish", "sea urchin", "sea cucumber", "cottontail rabbit", "hare", "Angora rabbit",
+                        "hamster", "porcupine", "fox squirrel", "marmot", "beaver", "guinea pig", "common sorrel horse",
+                        "zebra", "pig", "wild boar", "warthog", "hippopotamus", "ox", "water buffalo", "bison",
+                        "ram (adult male sheep)", "bighorn sheep", "Alpine ibex", "hartebeest", "impala (antelope)",
+                        "gazelle", "arabian camel", "llama", "weasel", "mink", "European polecat",
+                        "black-footed ferret", "otter", "skunk", "badger", "armadillo", "three-toed sloth", "orangutan",
+                        "gorilla", "chimpanzee", "gibbon", "siamang", "guenon", "patas monkey", "baboon", "macaque",
+                        "langur", "black-and-white colobus", "proboscis monkey", "marmoset", "white-headed capuchin",
+                        "howler monkey", "titi monkey", "Geoffroy's spider monkey", "common squirrel monkey",
+                        "ring-tailed lemur", "indri", "Asian elephant", "African bush elephant", "red panda",
+                        "giant panda", "snoek fish", "eel", "silver salmon", "rock beauty fish", "clownfish",
+                        "sturgeon", "gar fish", "lionfish", "pufferfish", "abacus", "abaya", "academic gown",
+                        "accordion", "acoustic guitar", "aircraft carrier", "airliner", "airship", "altar", "ambulance",
+                        "amphibious vehicle", "analog clock", "apiary", "apron", "trash can", "assault rifle",
+                        "backpack", "bakery", "balance beam", "balloon", "ballpoint pen", "Band-Aid", "banjo",
+                        "baluster / handrail", "barbell", "barber chair", "barbershop", "barn", "barometer", "barrel",
+                        "wheelbarrow", "baseball", "basketball", "bassinet", "bassoon", "swimming cap", "bath towel",
+                        "bathtub", "station wagon", "lighthouse", "beaker", "military hat (bearskin or shako)",
+                        "beer bottle", "beer glass", "bell tower", "baby bib", "tandem bicycle", "bikini",
+                        "ring binder", "binoculars", "birdhouse", "boathouse", "bobsleigh", "bolo tie", "poke bonnet",
+                        "bookcase", "bookstore", "bottle cap", "hunting bow", "bow tie", "brass memorial plaque", "bra",
+                        "breakwater", "breastplate", "broom", "bucket", "buckle", "bulletproof vest",
+                        "high-speed train", "butcher shop", "taxicab", "cauldron", "candle", "cannon", "canoe",
+                        "can opener", "cardigan", "car mirror", "carousel", "tool kit", "cardboard box / carton",
+                        "car wheel", "automated teller machine", "cassette", "cassette player", "castle", "catamaran",
+                        "CD player", "cello", "mobile phone", "chain", "chain-link fence", "chain mail", "chainsaw",
+                        "storage chest", "chiffonier", "bell or wind chime", "china cabinet", "Christmas stocking",
+                        "church", "movie theater", "cleaver", "cliff dwelling", "cloak", "clogs", "cocktail shaker",
+                        "coffee mug", "coffeemaker", "spiral or coil", "combination lock", "computer keyboard",
+                        "candy store", "container ship", "convertible", "corkscrew", "cornet", "cowboy boot",
+                        "cowboy hat", "cradle", "construction crane", "crash helmet", "crate", "infant bed",
+                        "Crock Pot", "croquet ball", "crutch", "cuirass", "dam", "desk", "desktop computer",
+                        "rotary dial telephone", "diaper", "digital clock", "digital watch", "dining table",
+                        "dishcloth", "dishwasher", "disc brake", "dock", "dog sled", "dome", "doormat", "drilling rig",
+                        "drum", "drumstick", "dumbbell", "Dutch oven", "electric fan", "electric guitar",
+                        "electric locomotive", "entertainment center", "envelope", "espresso machine", "face powder",
+                        "feather boa", "filing cabinet", "fireboat", "fire truck", "fire screen", "flagpole", "flute",
+                        "folding chair", "football helmet", "forklift", "fountain", "fountain pen", "four-poster bed",
+                        "freight car", "French horn", "frying pan", "fur coat", "garbage truck",
+                        "gas mask or respirator", "gas pump", "goblet", "go-kart", "golf ball", "golf cart", "gondola",
+                        "gong", "gown", "grand piano", "greenhouse", "radiator grille", "grocery store", "guillotine",
+                        "hair clip", "hair spray", "half-track", "hammer", "hamper", "hair dryer", "hand-held computer",
+                        "handkerchief", "hard disk drive", "harmonica", "harp", "combine harvester", "hatchet",
+                        "holster", "home theater", "honeycomb", "hook", "hoop skirt", "gymnastic horizontal bar",
+                        "horse-drawn vehicle", "hourglass", "iPod", "clothes iron", "carved pumpkin", "jeans", "jeep",
+                        "T-shirt", "jigsaw puzzle", "rickshaw", "joystick", "kimono", "knee pad", "knot", "lab coat",
+                        "ladle", "lampshade", "laptop computer", "lawn mower", "lens cap", "letter opener", "library",
+                        "lifeboat", "lighter", "limousine", "ocean liner", "lipstick", "slip-on shoe", "lotion",
+                        "music speaker", "loupe magnifying glass", "sawmill", "magnetic compass", "messenger bag",
+                        "mailbox", "tights", "one-piece bathing suit", "manhole cover", "maraca", "marimba", "mask",
+                        "matchstick", "maypole", "maze", "measuring cup", "medicine cabinet", "megalith", "microphone",
+                        "microwave oven", "military uniform", "milk can", "minibus", "miniskirt", "minivan", "missile",
+                        "mitten", "mixing bowl", "mobile home", "ford model t", "modem", "monastery", "monitor",
+                        "moped", "mortar and pestle", "graduation cap", "mosque", "mosquito net", "vespa",
+                        "mountain bike", "tent", "computer mouse", "mousetrap", "moving van", "muzzle", "metal nail",
+                        "neck brace", "necklace", "baby pacifier", "notebook computer", "obelisk", "oboe", "ocarina",
+                        "odometer", "oil filter", "pipe organ", "oscilloscope", "overskirt", "bullock cart",
+                        "oxygen mask", "product packet / packaging", "paddle", "paddle wheel", "padlock", "paintbrush",
+                        "pajamas", "palace", "pan flute", "paper towel", "parachute", "parallel bars", "park bench",
+                        "parking meter", "railroad car", "patio", "payphone", "pedestal", "pencil case",
+                        "pencil sharpener", "perfume", "Petri dish", "photocopier", "plectrum", "Pickelhaube",
+                        "picket fence", "pickup truck", "pier", "piggy bank", "pill bottle", "pillow", "ping-pong ball",
+                        "pinwheel", "pirate ship", "drink pitcher", "block plane", "planetarium", "plastic bag",
+                        "plate rack", "farm plow", "plunger", "Polaroid camera", "pole", "police van", "poncho",
+                        "pool table", "soda bottle", "plant pot", "potter's wheel", "power drill", "prayer rug",
+                        "printer", "prison", "missile", "projector", "hockey puck", "punching bag", "purse", "quill",
+                        "quilt", "race car", "racket", "radiator", "radio", "radio telescope", "rain barrel",
+                        "recreational vehicle", "fishing casting reel", "reflex camera", "refrigerator",
+                        "remote control", "restaurant", "revolver", "rifle", "rocking chair", "rotisserie", "eraser",
+                        "rugby ball", "ruler measuring stick", "sneaker", "safe", "safety pin", "salt shaker", "sandal",
+                        "sarong", "saxophone", "scabbard", "weighing scale", "school bus", "schooner", "scoreboard",
+                        "CRT monitor", "screw", "screwdriver", "seat belt", "sewing machine", "shield", "shoe store",
+                        "shoji screen / room divider", "shopping basket", "shopping cart", "shovel", "shower cap",
+                        "shower curtain", "ski", "balaclava ski mask", "sleeping bag", "slide rule", "sliding door",
+                        "slot machine", "snorkel", "snowmobile", "snowplow", "soap dispenser", "soccer ball", "sock",
+                        "solar thermal collector", "sombrero", "soup bowl", "keyboard space bar", "space heater",
+                        "space shuttle", "spatula", "motorboat", "spider web", "spindle", "sports car", "spotlight",
+                        "stage", "steam locomotive", "through arch bridge", "steel drum", "stethoscope", "scarf",
+                        "stone wall", "stopwatch", "stove", "strainer", "tram", "stretcher", "couch", "stupa",
+                        "submarine", "suit", "sundial", "sunglasses", "sunglasses", "sunscreen", "suspension bridge",
+                        "mop", "sweatshirt", "swim trunks / shorts", "swing", "electrical switch", "syringe",
+                        "table lamp", "tank", "tape player", "teapot", "teddy bear", "television", "tennis ball",
+                        "thatched roof", "front curtain", "thimble", "threshing machine", "throne", "tile roof",
+                        "toaster", "tobacco shop", "toilet seat", "torch", "totem pole", "tow truck", "toy store",
+                        "tractor", "semi-trailer truck", "tray", "trench coat", "tricycle", "trimaran", "tripod",
+                        "triumphal arch", "trolleybus", "trombone", "hot tub", "turnstile", "typewriter keyboard",
+                        "umbrella", "unicycle", "upright piano", "vacuum cleaner", "vase", "vaulted or arched ceiling",
+                        "velvet fabric", "vending machine", "vestment", "viaduct", "violin", "volleyball",
+                        "waffle iron", "wall clock", "wallet", "wardrobe", "military aircraft", "sink",
+                        "washing machine", "water bottle", "water jug", "water tower", "whiskey jug", "whistle",
+                        "hair wig", "window screen", "window shade", "Windsor tie", "wine bottle", "airplane wing",
+                        "wok", "wooden spoon", "wool", "split-rail fence", "shipwreck", "sailboat", "yurt", "website",
+                        "comic book", "crossword", "traffic or street sign", "traffic light", "dust jacket", "menu",
+                        "plate", "guacamole", "consomme", "hot pot", "trifle", "ice cream", "popsicle", "baguette",
+                        "bagel", "pretzel", "cheeseburger", "hot dog", "mashed potatoes", "cabbage", "broccoli",
+                        "cauliflower", "zucchini", "spaghetti squash", "acorn squash", "butternut squash", "cucumber",
+                        "artichoke", "bell pepper", "cardoon", "mushroom", "Granny Smith apple", "strawberry", "orange",
+                        "lemon", "fig", "pineapple", "banana", "jackfruit", "cherimoya (custard apple)", "pomegranate",
+                        "hay", "carbonara", "chocolate syrup", "dough", "meatloaf", "pizza", "pot pie", "burrito",
+                        "red wine", "espresso", "tea cup", "eggnog", "mountain", "bubble", "cliff", "coral reef",
+                        "geyser", "lakeshore", "promontory", "sandbar", "beach", "valley", "volcano", "baseball player",
+                        "bridegroom", "scuba diver", "rapeseed", "daisy", "yellow lady's slipper", "corn", "acorn",
+                        "rose hip", "horse chestnut seed", "coral fungus", "agaric", "gyromitra", "stinkhorn mushroom",
+                        "earth star fungus", "hen of the woods mushroom", "bolete", "corn cob", "toilet paper"]
+
+    # cifar100_classes = [name for name in os.listdir(args.data_path)
+    #                 if os.path.isdir(os.path.join(args.data_path, name))]
+
+    imagenet_single_template = [
+        'a photo of a {}.',
+    ]
+
+    imagenet_7_templates = [
+        'itap of a {}.',
+        'a origami {}.',
+        'a bad photo of the {}.',
+        'a photo of the large {}.',
+        'a {} in a video game.',
+        'art of the {}.',
+        'a photo of the small {}.',
+    ]
+
+
+    print('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    print('load data')
+    valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    print('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, imagenet_classes, imagenet_7_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    print(f'accuracy with text proxy: {top1:.2f}')
+
+    print('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    acc_lame = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+            beta = args.beta * math.sqrt((i + 1) / n)
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+            rho -= rlr * (tlabel - args.alpha / num_class)
+            rho[rho < 0] = 0
+            text_label[i, :] = tlabel
+            vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            grad = torch.outer(x, vision_label - tlabel)
+            w -= (lr / args.tau_i) * grad
+            w = F.normalize(w, dim=0)
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+
+        # MAPLS - EM Algorithm
+        pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+        qy = mapls(combo_label, pz=pz, qy_mode="soft", max_iter=100, lam=0.6)
+
+        w = np.array(qy) / np.array(pz)
+        if combo_label.is_cuda:
+            combo_label_cpu = combo_label.cpu()
+        qy_probs = lsc(combo_label_cpu, 1.0 / w)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        # LAME
+        # lame = LAME()
+        # qy_probs_device = qy_probs.to("cuda:1")
+        # image_feat_device = image_feat.to("cuda:1")
+        # lame_prob = lame.run_step(qy_probs_device, image_feat_device)
+        # acc1_lame, acc5_lame = accuracy(lame_prob, image_label[idx], topk=(1, 5))
+
+        lame = LAME()
+        batch_size = 64  # Set batch size for processing
+        num_batches = (combo_label.shape[0] + batch_size - 1) // batch_size  # Calculate number of batches
+
+        all_lame_probs = []
+        for batch_idx in range(num_batches):
+            start_idx = batch_idx * batch_size
+            end_idx = min((batch_idx + 1) * batch_size, combo_label.shape[0])
+
+            combo_label_batch = combo_label[start_idx:end_idx].to("cuda:0")
+            image_feat_batch = image_feat[start_idx:end_idx].to("cuda:0")
+
+            # Process each batch separately to avoid OOM
+            lame_prob = lame.run_step(combo_label_batch, image_feat_batch)
+            all_lame_probs.append(lame_prob)
+
+        # Concatenate the results from each batch
+        lame_prob = torch.cat(all_lame_probs, dim=0)
+
+        acc1_lame, acc5_lame = accuracy(lame_prob, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+        acc_lame[iter] = (acc1_lame / n) * 100
+    print('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    print('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    print('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+    print('mean acc of lame is: {:.2f}'.format(torch.mean(acc_lame)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    lams = [0.6]
+    for lam in lams:
+        main(lam)
+

OnZeta/main_online_imagenet_mapls_nonlinear.py

@@ -0,0 +1,368 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='/home/lt_test/projects/datasets/ImageNet/', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='RN50',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=512, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.8, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+
+class VisionProxyMLP(torch.nn.Module):
+    def __init__(self, input_dim, hidden_dim, output_dim):
+        super().__init__()
+        self.fc1 = torch.nn.Linear(input_dim, hidden_dim)
+        self.relu = torch.nn.ReLU()
+        self.fc2 = torch.nn.Linear(hidden_dim, output_dim)
+
+    def forward(self, x):
+        x = self.fc1(x)
+        x = self.relu(x)
+        x = self.fc2(x)
+        return x
+
+def main(lam):
+
+    args = parser.parse_args()
+    print(args)
+    imagenet_classes = ["tench", "goldfish", "great white shark", "tiger shark", "hammerhead shark", "electric ray",
+                        "stingray", "rooster", "hen", "ostrich", "brambling", "goldfinch", "house finch", "junco",
+                        "indigo bunting", "American robin", "bulbul", "jay", "magpie", "chickadee", "American dipper",
+                        "kite (bird of prey)", "bald eagle", "vulture", "great grey owl", "fire salamander",
+                        "smooth newt", "newt", "spotted salamander", "axolotl", "American bullfrog", "tree frog",
+                        "tailed frog", "loggerhead sea turtle", "leatherback sea turtle", "mud turtle", "terrapin",
+                        "box turtle", "banded gecko", "green iguana", "Carolina anole",
+                        "desert grassland whiptail lizard", "agama", "frilled-necked lizard", "alligator lizard",
+                        "Gila monster", "European green lizard", "chameleon", "Komodo dragon", "Nile crocodile",
+                        "American alligator", "triceratops", "worm snake", "ring-necked snake",
+                        "eastern hog-nosed snake", "smooth green snake", "kingsnake", "garter snake", "water snake",
+                        "vine snake", "night snake", "boa constrictor", "African rock python", "Indian cobra",
+                        "green mamba", "sea snake", "Saharan horned viper", "eastern diamondback rattlesnake",
+                        "sidewinder rattlesnake", "trilobite", "harvestman", "scorpion", "yellow garden spider",
+                        "barn spider", "European garden spider", "southern black widow", "tarantula", "wolf spider",
+                        "tick", "centipede", "black grouse", "ptarmigan", "ruffed grouse", "prairie grouse", "peafowl",
+                        "quail", "partridge", "african grey parrot", "macaw", "sulphur-crested cockatoo", "lorikeet",
+                        "coucal", "bee eater", "hornbill", "hummingbird", "jacamar", "toucan", "duck",
+                        "red-breasted merganser", "goose", "black swan", "tusker", "echidna", "platypus", "wallaby",
+                        "koala", "wombat", "jellyfish", "sea anemone", "brain coral", "flatworm", "nematode", "conch",
+                        "snail", "slug", "sea slug", "chiton", "chambered nautilus", "Dungeness crab", "rock crab",
+                        "fiddler crab", "red king crab", "American lobster", "spiny lobster", "crayfish", "hermit crab",
+                        "isopod", "white stork", "black stork", "spoonbill", "flamingo", "little blue heron",
+                        "great egret", "bittern bird", "crane bird", "limpkin", "common gallinule", "American coot",
+                        "bustard", "ruddy turnstone", "dunlin", "common redshank", "dowitcher", "oystercatcher",
+                        "pelican", "king penguin", "albatross", "grey whale", "killer whale", "dugong", "sea lion",
+                        "Chihuahua", "Japanese Chin", "Maltese", "Pekingese", "Shih Tzu", "King Charles Spaniel",
+                        "Papillon", "toy terrier", "Rhodesian Ridgeback", "Afghan Hound", "Basset Hound", "Beagle",
+                        "Bloodhound", "Bluetick Coonhound", "Black and Tan Coonhound", "Treeing Walker Coonhound",
+                        "English foxhound", "Redbone Coonhound", "borzoi", "Irish Wolfhound", "Italian Greyhound",
+                        "Whippet", "Ibizan Hound", "Norwegian Elkhound", "Otterhound", "Saluki", "Scottish Deerhound",
+                        "Weimaraner", "Staffordshire Bull Terrier", "American Staffordshire Terrier",
+                        "Bedlington Terrier", "Border Terrier", "Kerry Blue Terrier", "Irish Terrier",
+                        "Norfolk Terrier", "Norwich Terrier", "Yorkshire Terrier", "Wire Fox Terrier",
+                        "Lakeland Terrier", "Sealyham Terrier", "Airedale Terrier", "Cairn Terrier",
+                        "Australian Terrier", "Dandie Dinmont Terrier", "Boston Terrier", "Miniature Schnauzer",
+                        "Giant Schnauzer", "Standard Schnauzer", "Scottish Terrier", "Tibetan Terrier",
+                        "Australian Silky Terrier", "Soft-coated Wheaten Terrier", "West Highland White Terrier",
+                        "Lhasa Apso", "Flat-Coated Retriever", "Curly-coated Retriever", "Golden Retriever",
+                        "Labrador Retriever", "Chesapeake Bay Retriever", "German Shorthaired Pointer", "Vizsla",
+                        "English Setter", "Irish Setter", "Gordon Setter", "Brittany dog", "Clumber Spaniel",
+                        "English Springer Spaniel", "Welsh Springer Spaniel", "Cocker Spaniel", "Sussex Spaniel",
+                        "Irish Water Spaniel", "Kuvasz", "Schipperke", "Groenendael dog", "Malinois", "Briard",
+                        "Australian Kelpie", "Komondor", "Old English Sheepdog", "Shetland Sheepdog", "collie",
+                        "Border Collie", "Bouvier des Flandres dog", "Rottweiler", "German Shepherd Dog", "Dobermann",
+                        "Miniature Pinscher", "Greater Swiss Mountain Dog", "Bernese Mountain Dog",
+                        "Appenzeller Sennenhund", "Entlebucher Sennenhund", "Boxer", "Bullmastiff", "Tibetan Mastiff",
+                        "French Bulldog", "Great Dane", "St. Bernard", "husky", "Alaskan Malamute", "Siberian Husky",
+                        "Dalmatian", "Affenpinscher", "Basenji", "pug", "Leonberger", "Newfoundland dog",
+                        "Great Pyrenees dog", "Samoyed", "Pomeranian", "Chow Chow", "Keeshond", "brussels griffon",
+                        "Pembroke Welsh Corgi", "Cardigan Welsh Corgi", "Toy Poodle", "Miniature Poodle",
+                        "Standard Poodle", "Mexican hairless dog (xoloitzcuintli)", "grey wolf", "Alaskan tundra wolf",
+                        "red wolf or maned wolf", "coyote", "dingo", "dhole", "African wild dog", "hyena", "red fox",
+                        "kit fox", "Arctic fox", "grey fox", "tabby cat", "tiger cat", "Persian cat", "Siamese cat",
+                        "Egyptian Mau", "cougar", "lynx", "leopard", "snow leopard", "jaguar", "lion", "tiger",
+                        "cheetah", "brown bear", "American black bear", "polar bear", "sloth bear", "mongoose",
+                        "meerkat", "tiger beetle", "ladybug", "ground beetle", "longhorn beetle", "leaf beetle",
+                        "dung beetle", "rhinoceros beetle", "weevil", "fly", "bee", "ant", "grasshopper",
+                        "cricket insect", "stick insect", "cockroach", "praying mantis", "cicada", "leafhopper",
+                        "lacewing", "dragonfly", "damselfly", "red admiral butterfly", "ringlet butterfly",
+                        "monarch butterfly", "small white butterfly", "sulphur butterfly", "gossamer-winged butterfly",
+                        "starfish", "sea urchin", "sea cucumber", "cottontail rabbit", "hare", "Angora rabbit",
+                        "hamster", "porcupine", "fox squirrel", "marmot", "beaver", "guinea pig", "common sorrel horse",
+                        "zebra", "pig", "wild boar", "warthog", "hippopotamus", "ox", "water buffalo", "bison",
+                        "ram (adult male sheep)", "bighorn sheep", "Alpine ibex", "hartebeest", "impala (antelope)",
+                        "gazelle", "arabian camel", "llama", "weasel", "mink", "European polecat",
+                        "black-footed ferret", "otter", "skunk", "badger", "armadillo", "three-toed sloth", "orangutan",
+                        "gorilla", "chimpanzee", "gibbon", "siamang", "guenon", "patas monkey", "baboon", "macaque",
+                        "langur", "black-and-white colobus", "proboscis monkey", "marmoset", "white-headed capuchin",
+                        "howler monkey", "titi monkey", "Geoffroy's spider monkey", "common squirrel monkey",
+                        "ring-tailed lemur", "indri", "Asian elephant", "African bush elephant", "red panda",
+                        "giant panda", "snoek fish", "eel", "silver salmon", "rock beauty fish", "clownfish",
+                        "sturgeon", "gar fish", "lionfish", "pufferfish", "abacus", "abaya", "academic gown",
+                        "accordion", "acoustic guitar", "aircraft carrier", "airliner", "airship", "altar", "ambulance",
+                        "amphibious vehicle", "analog clock", "apiary", "apron", "trash can", "assault rifle",
+                        "backpack", "bakery", "balance beam", "balloon", "ballpoint pen", "Band-Aid", "banjo",
+                        "baluster / handrail", "barbell", "barber chair", "barbershop", "barn", "barometer", "barrel",
+                        "wheelbarrow", "baseball", "basketball", "bassinet", "bassoon", "swimming cap", "bath towel",
+                        "bathtub", "station wagon", "lighthouse", "beaker", "military hat (bearskin or shako)",
+                        "beer bottle", "beer glass", "bell tower", "baby bib", "tandem bicycle", "bikini",
+                        "ring binder", "binoculars", "birdhouse", "boathouse", "bobsleigh", "bolo tie", "poke bonnet",
+                        "bookcase", "bookstore", "bottle cap", "hunting bow", "bow tie", "brass memorial plaque", "bra",
+                        "breakwater", "breastplate", "broom", "bucket", "buckle", "bulletproof vest",
+                        "high-speed train", "butcher shop", "taxicab", "cauldron", "candle", "cannon", "canoe",
+                        "can opener", "cardigan", "car mirror", "carousel", "tool kit", "cardboard box / carton",
+                        "car wheel", "automated teller machine", "cassette", "cassette player", "castle", "catamaran",
+                        "CD player", "cello", "mobile phone", "chain", "chain-link fence", "chain mail", "chainsaw",
+                        "storage chest", "chiffonier", "bell or wind chime", "china cabinet", "Christmas stocking",
+                        "church", "movie theater", "cleaver", "cliff dwelling", "cloak", "clogs", "cocktail shaker",
+                        "coffee mug", "coffeemaker", "spiral or coil", "combination lock", "computer keyboard",
+                        "candy store", "container ship", "convertible", "corkscrew", "cornet", "cowboy boot",
+                        "cowboy hat", "cradle", "construction crane", "crash helmet", "crate", "infant bed",
+                        "Crock Pot", "croquet ball", "crutch", "cuirass", "dam", "desk", "desktop computer",
+                        "rotary dial telephone", "diaper", "digital clock", "digital watch", "dining table",
+                        "dishcloth", "dishwasher", "disc brake", "dock", "dog sled", "dome", "doormat", "drilling rig",
+                        "drum", "drumstick", "dumbbell", "Dutch oven", "electric fan", "electric guitar",
+                        "electric locomotive", "entertainment center", "envelope", "espresso machine", "face powder",
+                        "feather boa", "filing cabinet", "fireboat", "fire truck", "fire screen", "flagpole", "flute",
+                        "folding chair", "football helmet", "forklift", "fountain", "fountain pen", "four-poster bed",
+                        "freight car", "French horn", "frying pan", "fur coat", "garbage truck",
+                        "gas mask or respirator", "gas pump", "goblet", "go-kart", "golf ball", "golf cart", "gondola",
+                        "gong", "gown", "grand piano", "greenhouse", "radiator grille", "grocery store", "guillotine",
+                        "hair clip", "hair spray", "half-track", "hammer", "hamper", "hair dryer", "hand-held computer",
+                        "handkerchief", "hard disk drive", "harmonica", "harp", "combine harvester", "hatchet",
+                        "holster", "home theater", "honeycomb", "hook", "hoop skirt", "gymnastic horizontal bar",
+                        "horse-drawn vehicle", "hourglass", "iPod", "clothes iron", "carved pumpkin", "jeans", "jeep",
+                        "T-shirt", "jigsaw puzzle", "rickshaw", "joystick", "kimono", "knee pad", "knot", "lab coat",
+                        "ladle", "lampshade", "laptop computer", "lawn mower", "lens cap", "letter opener", "library",
+                        "lifeboat", "lighter", "limousine", "ocean liner", "lipstick", "slip-on shoe", "lotion",
+                        "music speaker", "loupe magnifying glass", "sawmill", "magnetic compass", "messenger bag",
+                        "mailbox", "tights", "one-piece bathing suit", "manhole cover", "maraca", "marimba", "mask",
+                        "matchstick", "maypole", "maze", "measuring cup", "medicine cabinet", "megalith", "microphone",
+                        "microwave oven", "military uniform", "milk can", "minibus", "miniskirt", "minivan", "missile",
+                        "mitten", "mixing bowl", "mobile home", "ford model t", "modem", "monastery", "monitor",
+                        "moped", "mortar and pestle", "graduation cap", "mosque", "mosquito net", "vespa",
+                        "mountain bike", "tent", "computer mouse", "mousetrap", "moving van", "muzzle", "metal nail",
+                        "neck brace", "necklace", "baby pacifier", "notebook computer", "obelisk", "oboe", "ocarina",
+                        "odometer", "oil filter", "pipe organ", "oscilloscope", "overskirt", "bullock cart",
+                        "oxygen mask", "product packet / packaging", "paddle", "paddle wheel", "padlock", "paintbrush",
+                        "pajamas", "palace", "pan flute", "paper towel", "parachute", "parallel bars", "park bench",
+                        "parking meter", "railroad car", "patio", "payphone", "pedestal", "pencil case",
+                        "pencil sharpener", "perfume", "Petri dish", "photocopier", "plectrum", "Pickelhaube",
+                        "picket fence", "pickup truck", "pier", "piggy bank", "pill bottle", "pillow", "ping-pong ball",
+                        "pinwheel", "pirate ship", "drink pitcher", "block plane", "planetarium", "plastic bag",
+                        "plate rack", "farm plow", "plunger", "Polaroid camera", "pole", "police van", "poncho",
+                        "pool table", "soda bottle", "plant pot", "potter's wheel", "power drill", "prayer rug",
+                        "printer", "prison", "missile", "projector", "hockey puck", "punching bag", "purse", "quill",
+                        "quilt", "race car", "racket", "radiator", "radio", "radio telescope", "rain barrel",
+                        "recreational vehicle", "fishing casting reel", "reflex camera", "refrigerator",
+                        "remote control", "restaurant", "revolver", "rifle", "rocking chair", "rotisserie", "eraser",
+                        "rugby ball", "ruler measuring stick", "sneaker", "safe", "safety pin", "salt shaker", "sandal",
+                        "sarong", "saxophone", "scabbard", "weighing scale", "school bus", "schooner", "scoreboard",
+                        "CRT monitor", "screw", "screwdriver", "seat belt", "sewing machine", "shield", "shoe store",
+                        "shoji screen / room divider", "shopping basket", "shopping cart", "shovel", "shower cap",
+                        "shower curtain", "ski", "balaclava ski mask", "sleeping bag", "slide rule", "sliding door",
+                        "slot machine", "snorkel", "snowmobile", "snowplow", "soap dispenser", "soccer ball", "sock",
+                        "solar thermal collector", "sombrero", "soup bowl", "keyboard space bar", "space heater",
+                        "space shuttle", "spatula", "motorboat", "spider web", "spindle", "sports car", "spotlight",
+                        "stage", "steam locomotive", "through arch bridge", "steel drum", "stethoscope", "scarf",
+                        "stone wall", "stopwatch", "stove", "strainer", "tram", "stretcher", "couch", "stupa",
+                        "submarine", "suit", "sundial", "sunglasses", "sunglasses", "sunscreen", "suspension bridge",
+                        "mop", "sweatshirt", "swim trunks / shorts", "swing", "electrical switch", "syringe",
+                        "table lamp", "tank", "tape player", "teapot", "teddy bear", "television", "tennis ball",
+                        "thatched roof", "front curtain", "thimble", "threshing machine", "throne", "tile roof",
+                        "toaster", "tobacco shop", "toilet seat", "torch", "totem pole", "tow truck", "toy store",
+                        "tractor", "semi-trailer truck", "tray", "trench coat", "tricycle", "trimaran", "tripod",
+                        "triumphal arch", "trolleybus", "trombone", "hot tub", "turnstile", "typewriter keyboard",
+                        "umbrella", "unicycle", "upright piano", "vacuum cleaner", "vase", "vaulted or arched ceiling",
+                        "velvet fabric", "vending machine", "vestment", "viaduct", "violin", "volleyball",
+                        "waffle iron", "wall clock", "wallet", "wardrobe", "military aircraft", "sink",
+                        "washing machine", "water bottle", "water jug", "water tower", "whiskey jug", "whistle",
+                        "hair wig", "window screen", "window shade", "Windsor tie", "wine bottle", "airplane wing",
+                        "wok", "wooden spoon", "wool", "split-rail fence", "shipwreck", "sailboat", "yurt", "website",
+                        "comic book", "crossword", "traffic or street sign", "traffic light", "dust jacket", "menu",
+                        "plate", "guacamole", "consomme", "hot pot", "trifle", "ice cream", "popsicle", "baguette",
+                        "bagel", "pretzel", "cheeseburger", "hot dog", "mashed potatoes", "cabbage", "broccoli",
+                        "cauliflower", "zucchini", "spaghetti squash", "acorn squash", "butternut squash", "cucumber",
+                        "artichoke", "bell pepper", "cardoon", "mushroom", "Granny Smith apple", "strawberry", "orange",
+                        "lemon", "fig", "pineapple", "banana", "jackfruit", "cherimoya (custard apple)", "pomegranate",
+                        "hay", "carbonara", "chocolate syrup", "dough", "meatloaf", "pizza", "pot pie", "burrito",
+                        "red wine", "espresso", "tea cup", "eggnog", "mountain", "bubble", "cliff", "coral reef",
+                        "geyser", "lakeshore", "promontory", "sandbar", "beach", "valley", "volcano", "baseball player",
+                        "bridegroom", "scuba diver", "rapeseed", "daisy", "yellow lady's slipper", "corn", "acorn",
+                        "rose hip", "horse chestnut seed", "coral fungus", "agaric", "gyromitra", "stinkhorn mushroom",
+                        "earth star fungus", "hen of the woods mushroom", "bolete", "corn cob", "toilet paper"]
+
+    # cifar100_classes = [name for name in os.listdir(args.data_path)
+    #                 if os.path.isdir(os.path.join(args.data_path, name))]
+
+    imagenet_single_template = [
+        'a photo of a {}.',
+    ]
+
+    imagenet_7_templates = [
+        'itap of a {}.',
+        'a origami {}.',
+        'a bad photo of the {}.',
+        'a photo of the large {}.',
+        'a {} in a video game.',
+        'art of the {}.',
+        'a photo of the small {}.',
+    ]
+
+
+    print('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    print('load data')
+    valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    print('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, imagenet_classes, imagenet_7_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    print(f'accuracy with text proxy: {top1:.2f}')
+
+    print('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    feat_dim = image_feat.shape[1]
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        mlp = VisionProxyMLP(feat_dim, hidden_dim=512, output_dim=num_class).cuda()
+        optimizer = torch.optim.SGD(mlp.parameters(), lr=args.cw)
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+
+            for param_group in optimizer.param_groups:
+                param_group['lr'] = lr
+
+            beta = args.beta * math.sqrt((i + 1) / n)
+
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+            rho -= rlr * (tlabel - args.alpha / num_class)
+            rho[rho < 0] = 0
+            text_label[i, :] = tlabel
+
+
+            mlp.train()
+            vision_logits = mlp(x) / args.tau_i  # [1, num_class]
+            vision_label = F.softmax(vision_logits, dim=0)
+
+            # --- 组合标签 ---
+            combo_label[i, :] = (beta * vision_label + (1 - beta) * tlabel).detach()
+
+            # --- 计算 loss 并优化 MLP ---
+            # grad_vector = torch.outer(x, vision_label - tlabel)  # shape: [1, C]
+            # vision_logits.backward(gradient=grad_vector)
+            grad_vector = vision_label - tlabel  # shape: [1000]
+            vision_logits.backward(gradient=grad_vector)
+            optimizer.step()
+            # vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            # combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            # grad = torch.outer(x, vision_label - tlabel)
+            # w -= (lr / args.tau_i) * grad
+            # w = F.normalize(w, dim=0)
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+
+        # MAPLS - EM Algorithm
+        pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+        qy = mapls(combo_label, pz = pz, qy_mode = "soft", max_iter = 100, lam = lam)
+
+        w = np.array(qy) / np.array(pz)
+        if combo_label.is_cuda:
+            combo_label_cpu = combo_label.cpu()
+        qy_probs = lsc(combo_label_cpu, 1.0/w)
+        qy_probs = torch.from_numpy(qy_probs)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+    print('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    print('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    print('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    lams = [0.6]
+    for lam in lams:
+        main(lam)
+

OnZeta/main_online_imagenet_margin_softmax.py

@@ -0,0 +1,346 @@
+# Copyright (c) Alibaba Group
+import argparse
+import torch
+import torchvision.datasets as datasets
+import torch.nn.functional as F
+import clip
+import os
+import math
+import numpy as np
+
+from MAPLS.mapls import mapls
+from MAPLS.common import lsc
+
+model_names = ['RN50', 'ViT-B/32', 'ViT-B/16', 'ViT-L/14', 'ViT-L/14@336px']
+parser = argparse.ArgumentParser(description='OnZeta for ImageNet')
+parser.add_argument('--data_path', default='/home/lt_test/projects/datasets/ImageNet/', type=str,
+                    help='dataset path')
+parser.add_argument('-a', '--arch', metavar='ARCH', default='ViT-B/16',
+                    choices=model_names,
+                    help='model architecture: ' +
+                         ' | '.join(model_names) +
+                         ' (default: RN50)')
+parser.add_argument('-j', '--workers', default=8, type=int, metavar='N',
+                    help='number of data loading workers (default: 8)')
+parser.add_argument('-b', '--batch-size', default=256, type=int,
+                    metavar='N',
+                    help='mini-batch size (default: 256)')
+parser.add_argument('--tau_t', default=0.01, type=float)
+parser.add_argument('--tau_i', default=0.04, type=float)
+parser.add_argument('--cw', default=0.5, type=float)
+parser.add_argument('--cr', default=20, type=float)
+parser.add_argument('--alpha', default=1, type=float)
+parser.add_argument('--beta', default=0.8, type=float)
+parser.add_argument('--repeat', default=5, type=int)
+
+parser.add_argument('--lam', default=0.6, type=float)
+parser.add_argument('--margin', default=0.2, type=float, help='margin for margin softmax')
+
+
+def main(margin):
+
+    args = parser.parse_args()
+    print(args)
+    imagenet_classes = ["tench", "goldfish", "great white shark", "tiger shark", "hammerhead shark", "electric ray",
+                        "stingray", "rooster", "hen", "ostrich", "brambling", "goldfinch", "house finch", "junco",
+                        "indigo bunting", "American robin", "bulbul", "jay", "magpie", "chickadee", "American dipper",
+                        "kite (bird of prey)", "bald eagle", "vulture", "great grey owl", "fire salamander",
+                        "smooth newt", "newt", "spotted salamander", "axolotl", "American bullfrog", "tree frog",
+                        "tailed frog", "loggerhead sea turtle", "leatherback sea turtle", "mud turtle", "terrapin",
+                        "box turtle", "banded gecko", "green iguana", "Carolina anole",
+                        "desert grassland whiptail lizard", "agama", "frilled-necked lizard", "alligator lizard",
+                        "Gila monster", "European green lizard", "chameleon", "Komodo dragon", "Nile crocodile",
+                        "American alligator", "triceratops", "worm snake", "ring-necked snake",
+                        "eastern hog-nosed snake", "smooth green snake", "kingsnake", "garter snake", "water snake",
+                        "vine snake", "night snake", "boa constrictor", "African rock python", "Indian cobra",
+                        "green mamba", "sea snake", "Saharan horned viper", "eastern diamondback rattlesnake",
+                        "sidewinder rattlesnake", "trilobite", "harvestman", "scorpion", "yellow garden spider",
+                        "barn spider", "European garden spider", "southern black widow", "tarantula", "wolf spider",
+                        "tick", "centipede", "black grouse", "ptarmigan", "ruffed grouse", "prairie grouse", "peafowl",
+                        "quail", "partridge", "african grey parrot", "macaw", "sulphur-crested cockatoo", "lorikeet",
+                        "coucal", "bee eater", "hornbill", "hummingbird", "jacamar", "toucan", "duck",
+                        "red-breasted merganser", "goose", "black swan", "tusker", "echidna", "platypus", "wallaby",
+                        "koala", "wombat", "jellyfish", "sea anemone", "brain coral", "flatworm", "nematode", "conch",
+                        "snail", "slug", "sea slug", "chiton", "chambered nautilus", "Dungeness crab", "rock crab",
+                        "fiddler crab", "red king crab", "American lobster", "spiny lobster", "crayfish", "hermit crab",
+                        "isopod", "white stork", "black stork", "spoonbill", "flamingo", "little blue heron",
+                        "great egret", "bittern bird", "crane bird", "limpkin", "common gallinule", "American coot",
+                        "bustard", "ruddy turnstone", "dunlin", "common redshank", "dowitcher", "oystercatcher",
+                        "pelican", "king penguin", "albatross", "grey whale", "killer whale", "dugong", "sea lion",
+                        "Chihuahua", "Japanese Chin", "Maltese", "Pekingese", "Shih Tzu", "King Charles Spaniel",
+                        "Papillon", "toy terrier", "Rhodesian Ridgeback", "Afghan Hound", "Basset Hound", "Beagle",
+                        "Bloodhound", "Bluetick Coonhound", "Black and Tan Coonhound", "Treeing Walker Coonhound",
+                        "English foxhound", "Redbone Coonhound", "borzoi", "Irish Wolfhound", "Italian Greyhound",
+                        "Whippet", "Ibizan Hound", "Norwegian Elkhound", "Otterhound", "Saluki", "Scottish Deerhound",
+                        "Weimaraner", "Staffordshire Bull Terrier", "American Staffordshire Terrier",
+                        "Bedlington Terrier", "Border Terrier", "Kerry Blue Terrier", "Irish Terrier",
+                        "Norfolk Terrier", "Norwich Terrier", "Yorkshire Terrier", "Wire Fox Terrier",
+                        "Lakeland Terrier", "Sealyham Terrier", "Airedale Terrier", "Cairn Terrier",
+                        "Australian Terrier", "Dandie Dinmont Terrier", "Boston Terrier", "Miniature Schnauzer",
+                        "Giant Schnauzer", "Standard Schnauzer", "Scottish Terrier", "Tibetan Terrier",
+                        "Australian Silky Terrier", "Soft-coated Wheaten Terrier", "West Highland White Terrier",
+                        "Lhasa Apso", "Flat-Coated Retriever", "Curly-coated Retriever", "Golden Retriever",
+                        "Labrador Retriever", "Chesapeake Bay Retriever", "German Shorthaired Pointer", "Vizsla",
+                        "English Setter", "Irish Setter", "Gordon Setter", "Brittany dog", "Clumber Spaniel",
+                        "English Springer Spaniel", "Welsh Springer Spaniel", "Cocker Spaniel", "Sussex Spaniel",
+                        "Irish Water Spaniel", "Kuvasz", "Schipperke", "Groenendael dog", "Malinois", "Briard",
+                        "Australian Kelpie", "Komondor", "Old English Sheepdog", "Shetland Sheepdog", "collie",
+                        "Border Collie", "Bouvier des Flandres dog", "Rottweiler", "German Shepherd Dog", "Dobermann",
+                        "Miniature Pinscher", "Greater Swiss Mountain Dog", "Bernese Mountain Dog",
+                        "Appenzeller Sennenhund", "Entlebucher Sennenhund", "Boxer", "Bullmastiff", "Tibetan Mastiff",
+                        "French Bulldog", "Great Dane", "St. Bernard", "husky", "Alaskan Malamute", "Siberian Husky",
+                        "Dalmatian", "Affenpinscher", "Basenji", "pug", "Leonberger", "Newfoundland dog",
+                        "Great Pyrenees dog", "Samoyed", "Pomeranian", "Chow Chow", "Keeshond", "brussels griffon",
+                        "Pembroke Welsh Corgi", "Cardigan Welsh Corgi", "Toy Poodle", "Miniature Poodle",
+                        "Standard Poodle", "Mexican hairless dog (xoloitzcuintli)", "grey wolf", "Alaskan tundra wolf",
+                        "red wolf or maned wolf", "coyote", "dingo", "dhole", "African wild dog", "hyena", "red fox",
+                        "kit fox", "Arctic fox", "grey fox", "tabby cat", "tiger cat", "Persian cat", "Siamese cat",
+                        "Egyptian Mau", "cougar", "lynx", "leopard", "snow leopard", "jaguar", "lion", "tiger",
+                        "cheetah", "brown bear", "American black bear", "polar bear", "sloth bear", "mongoose",
+                        "meerkat", "tiger beetle", "ladybug", "ground beetle", "longhorn beetle", "leaf beetle",
+                        "dung beetle", "rhinoceros beetle", "weevil", "fly", "bee", "ant", "grasshopper",
+                        "cricket insect", "stick insect", "cockroach", "praying mantis", "cicada", "leafhopper",
+                        "lacewing", "dragonfly", "damselfly", "red admiral butterfly", "ringlet butterfly",
+                        "monarch butterfly", "small white butterfly", "sulphur butterfly", "gossamer-winged butterfly",
+                        "starfish", "sea urchin", "sea cucumber", "cottontail rabbit", "hare", "Angora rabbit",
+                        "hamster", "porcupine", "fox squirrel", "marmot", "beaver", "guinea pig", "common sorrel horse",
+                        "zebra", "pig", "wild boar", "warthog", "hippopotamus", "ox", "water buffalo", "bison",
+                        "ram (adult male sheep)", "bighorn sheep", "Alpine ibex", "hartebeest", "impala (antelope)",
+                        "gazelle", "arabian camel", "llama", "weasel", "mink", "European polecat",
+                        "black-footed ferret", "otter", "skunk", "badger", "armadillo", "three-toed sloth", "orangutan",
+                        "gorilla", "chimpanzee", "gibbon", "siamang", "guenon", "patas monkey", "baboon", "macaque",
+                        "langur", "black-and-white colobus", "proboscis monkey", "marmoset", "white-headed capuchin",
+                        "howler monkey", "titi monkey", "Geoffroy's spider monkey", "common squirrel monkey",
+                        "ring-tailed lemur", "indri", "Asian elephant", "African bush elephant", "red panda",
+                        "giant panda", "snoek fish", "eel", "silver salmon", "rock beauty fish", "clownfish",
+                        "sturgeon", "gar fish", "lionfish", "pufferfish", "abacus", "abaya", "academic gown",
+                        "accordion", "acoustic guitar", "aircraft carrier", "airliner", "airship", "altar", "ambulance",
+                        "amphibious vehicle", "analog clock", "apiary", "apron", "trash can", "assault rifle",
+                        "backpack", "bakery", "balance beam", "balloon", "ballpoint pen", "Band-Aid", "banjo",
+                        "baluster / handrail", "barbell", "barber chair", "barbershop", "barn", "barometer", "barrel",
+                        "wheelbarrow", "baseball", "basketball", "bassinet", "bassoon", "swimming cap", "bath towel",
+                        "bathtub", "station wagon", "lighthouse", "beaker", "military hat (bearskin or shako)",
+                        "beer bottle", "beer glass", "bell tower", "baby bib", "tandem bicycle", "bikini",
+                        "ring binder", "binoculars", "birdhouse", "boathouse", "bobsleigh", "bolo tie", "poke bonnet",
+                        "bookcase", "bookstore", "bottle cap", "hunting bow", "bow tie", "brass memorial plaque", "bra",
+                        "breakwater", "breastplate", "broom", "bucket", "buckle", "bulletproof vest",
+                        "high-speed train", "butcher shop", "taxicab", "cauldron", "candle", "cannon", "canoe",
+                        "can opener", "cardigan", "car mirror", "carousel", "tool kit", "cardboard box / carton",
+                        "car wheel", "automated teller machine", "cassette", "cassette player", "castle", "catamaran",
+                        "CD player", "cello", "mobile phone", "chain", "chain-link fence", "chain mail", "chainsaw",
+                        "storage chest", "chiffonier", "bell or wind chime", "china cabinet", "Christmas stocking",
+                        "church", "movie theater", "cleaver", "cliff dwelling", "cloak", "clogs", "cocktail shaker",
+                        "coffee mug", "coffeemaker", "spiral or coil", "combination lock", "computer keyboard",
+                        "candy store", "container ship", "convertible", "corkscrew", "cornet", "cowboy boot",
+                        "cowboy hat", "cradle", "construction crane", "crash helmet", "crate", "infant bed",
+                        "Crock Pot", "croquet ball", "crutch", "cuirass", "dam", "desk", "desktop computer",
+                        "rotary dial telephone", "diaper", "digital clock", "digital watch", "dining table",
+                        "dishcloth", "dishwasher", "disc brake", "dock", "dog sled", "dome", "doormat", "drilling rig",
+                        "drum", "drumstick", "dumbbell", "Dutch oven", "electric fan", "electric guitar",
+                        "electric locomotive", "entertainment center", "envelope", "espresso machine", "face powder",
+                        "feather boa", "filing cabinet", "fireboat", "fire truck", "fire screen", "flagpole", "flute",
+                        "folding chair", "football helmet", "forklift", "fountain", "fountain pen", "four-poster bed",
+                        "freight car", "French horn", "frying pan", "fur coat", "garbage truck",
+                        "gas mask or respirator", "gas pump", "goblet", "go-kart", "golf ball", "golf cart", "gondola",
+                        "gong", "gown", "grand piano", "greenhouse", "radiator grille", "grocery store", "guillotine",
+                        "hair clip", "hair spray", "half-track", "hammer", "hamper", "hair dryer", "hand-held computer",
+                        "handkerchief", "hard disk drive", "harmonica", "harp", "combine harvester", "hatchet",
+                        "holster", "home theater", "honeycomb", "hook", "hoop skirt", "gymnastic horizontal bar",
+                        "horse-drawn vehicle", "hourglass", "iPod", "clothes iron", "carved pumpkin", "jeans", "jeep",
+                        "T-shirt", "jigsaw puzzle", "rickshaw", "joystick", "kimono", "knee pad", "knot", "lab coat",
+                        "ladle", "lampshade", "laptop computer", "lawn mower", "lens cap", "letter opener", "library",
+                        "lifeboat", "lighter", "limousine", "ocean liner", "lipstick", "slip-on shoe", "lotion",
+                        "music speaker", "loupe magnifying glass", "sawmill", "magnetic compass", "messenger bag",
+                        "mailbox", "tights", "one-piece bathing suit", "manhole cover", "maraca", "marimba", "mask",
+                        "matchstick", "maypole", "maze", "measuring cup", "medicine cabinet", "megalith", "microphone",
+                        "microwave oven", "military uniform", "milk can", "minibus", "miniskirt", "minivan", "missile",
+                        "mitten", "mixing bowl", "mobile home", "ford model t", "modem", "monastery", "monitor",
+                        "moped", "mortar and pestle", "graduation cap", "mosque", "mosquito net", "vespa",
+                        "mountain bike", "tent", "computer mouse", "mousetrap", "moving van", "muzzle", "metal nail",
+                        "neck brace", "necklace", "baby pacifier", "notebook computer", "obelisk", "oboe", "ocarina",
+                        "odometer", "oil filter", "pipe organ", "oscilloscope", "overskirt", "bullock cart",
+                        "oxygen mask", "product packet / packaging", "paddle", "paddle wheel", "padlock", "paintbrush",
+                        "pajamas", "palace", "pan flute", "paper towel", "parachute", "parallel bars", "park bench",
+                        "parking meter", "railroad car", "patio", "payphone", "pedestal", "pencil case",
+                        "pencil sharpener", "perfume", "Petri dish", "photocopier", "plectrum", "Pickelhaube",
+                        "picket fence", "pickup truck", "pier", "piggy bank", "pill bottle", "pillow", "ping-pong ball",
+                        "pinwheel", "pirate ship", "drink pitcher", "block plane", "planetarium", "plastic bag",
+                        "plate rack", "farm plow", "plunger", "Polaroid camera", "pole", "police van", "poncho",
+                        "pool table", "soda bottle", "plant pot", "potter's wheel", "power drill", "prayer rug",
+                        "printer", "prison", "missile", "projector", "hockey puck", "punching bag", "purse", "quill",
+                        "quilt", "race car", "racket", "radiator", "radio", "radio telescope", "rain barrel",
+                        "recreational vehicle", "fishing casting reel", "reflex camera", "refrigerator",
+                        "remote control", "restaurant", "revolver", "rifle", "rocking chair", "rotisserie", "eraser",
+                        "rugby ball", "ruler measuring stick", "sneaker", "safe", "safety pin", "salt shaker", "sandal",
+                        "sarong", "saxophone", "scabbard", "weighing scale", "school bus", "schooner", "scoreboard",
+                        "CRT monitor", "screw", "screwdriver", "seat belt", "sewing machine", "shield", "shoe store",
+                        "shoji screen / room divider", "shopping basket", "shopping cart", "shovel", "shower cap",
+                        "shower curtain", "ski", "balaclava ski mask", "sleeping bag", "slide rule", "sliding door",
+                        "slot machine", "snorkel", "snowmobile", "snowplow", "soap dispenser", "soccer ball", "sock",
+                        "solar thermal collector", "sombrero", "soup bowl", "keyboard space bar", "space heater",
+                        "space shuttle", "spatula", "motorboat", "spider web", "spindle", "sports car", "spotlight",
+                        "stage", "steam locomotive", "through arch bridge", "steel drum", "stethoscope", "scarf",
+                        "stone wall", "stopwatch", "stove", "strainer", "tram", "stretcher", "couch", "stupa",
+                        "submarine", "suit", "sundial", "sunglasses", "sunglasses", "sunscreen", "suspension bridge",
+                        "mop", "sweatshirt", "swim trunks / shorts", "swing", "electrical switch", "syringe",
+                        "table lamp", "tank", "tape player", "teapot", "teddy bear", "television", "tennis ball",
+                        "thatched roof", "front curtain", "thimble", "threshing machine", "throne", "tile roof",
+                        "toaster", "tobacco shop", "toilet seat", "torch", "totem pole", "tow truck", "toy store",
+                        "tractor", "semi-trailer truck", "tray", "trench coat", "tricycle", "trimaran", "tripod",
+                        "triumphal arch", "trolleybus", "trombone", "hot tub", "turnstile", "typewriter keyboard",
+                        "umbrella", "unicycle", "upright piano", "vacuum cleaner", "vase", "vaulted or arched ceiling",
+                        "velvet fabric", "vending machine", "vestment", "viaduct", "violin", "volleyball",
+                        "waffle iron", "wall clock", "wallet", "wardrobe", "military aircraft", "sink",
+                        "washing machine", "water bottle", "water jug", "water tower", "whiskey jug", "whistle",
+                        "hair wig", "window screen", "window shade", "Windsor tie", "wine bottle", "airplane wing",
+                        "wok", "wooden spoon", "wool", "split-rail fence", "shipwreck", "sailboat", "yurt", "website",
+                        "comic book", "crossword", "traffic or street sign", "traffic light", "dust jacket", "menu",
+                        "plate", "guacamole", "consomme", "hot pot", "trifle", "ice cream", "popsicle", "baguette",
+                        "bagel", "pretzel", "cheeseburger", "hot dog", "mashed potatoes", "cabbage", "broccoli",
+                        "cauliflower", "zucchini", "spaghetti squash", "acorn squash", "butternut squash", "cucumber",
+                        "artichoke", "bell pepper", "cardoon", "mushroom", "Granny Smith apple", "strawberry", "orange",
+                        "lemon", "fig", "pineapple", "banana", "jackfruit", "cherimoya (custard apple)", "pomegranate",
+                        "hay", "carbonara", "chocolate syrup", "dough", "meatloaf", "pizza", "pot pie", "burrito",
+                        "red wine", "espresso", "tea cup", "eggnog", "mountain", "bubble", "cliff", "coral reef",
+                        "geyser", "lakeshore", "promontory", "sandbar", "beach", "valley", "volcano", "baseball player",
+                        "bridegroom", "scuba diver", "rapeseed", "daisy", "yellow lady's slipper", "corn", "acorn",
+                        "rose hip", "horse chestnut seed", "coral fungus", "agaric", "gyromitra", "stinkhorn mushroom",
+                        "earth star fungus", "hen of the woods mushroom", "bolete", "corn cob", "toilet paper"]
+
+    # cifar100_classes = [name for name in os.listdir(args.data_path)
+    #                 if os.path.isdir(os.path.join(args.data_path, name))]
+
+    imagenet_single_template = [
+        'a photo of a {}.',
+    ]
+
+    imagenet_7_templates = [
+        'itap of a {}.',
+        'a origami {}.',
+        'a bad photo of the {}.',
+        'a photo of the large {}.',
+        'a {} in a video game.',
+        'art of the {}.',
+        'a photo of the small {}.',
+    ]
+
+
+    print('load pre-trained model')
+    model, preprocess = clip.load(args.arch)
+    model = model.cuda()
+    model.eval()
+
+    print('load data')
+    print("margin is ",margin)
+    valdir = os.path.join(args.data_path, 'val')
+    # valdir = os.path.join(args.data_path, '')
+    val_set = datasets.ImageFolder(valdir, transform=preprocess)
+    loader = torch.utils.data.DataLoader(val_set, batch_size=args.batch_size, num_workers=args.workers)
+    with torch.no_grad():
+        image_feat = []
+        image_label = []
+        for i, (images, target) in enumerate(loader):
+            images = images.cuda()
+            target = target.cuda()
+            image_features = model.encode_image(images)
+            image_feat.append(F.normalize(image_features, dim=1))
+            image_label.append(target)
+    image_feat = torch.cat(image_feat, dim=0)
+    image_label = torch.cat(image_label, dim=0)
+    n = len(image_label)
+    image_feat = image_feat.float()
+
+    print('obtain text proxy')
+    text_classifier = zeroshot_classifier(clip, model, imagenet_classes, imagenet_7_templates)
+    text_classifier = text_classifier.float()
+    logits_t = image_feat @ text_classifier
+    acc1, acc5 = accuracy(logits_t, image_label, topk=(1, 5))
+    top1 = (acc1 / n) * 100
+    print(f'accuracy with text proxy: {top1:.2f}')
+
+    print('online zero-shot transfer: repeat {} times'.format(args.repeat))
+    num_class = len(torch.unique(image_label))
+    acc_onzeta = torch.zeros(args.repeat).cuda()
+    acc_onlab = torch.zeros(args.repeat).cuda()
+    acc_ls = torch.zeros(args.repeat).cuda()
+    for iter in range(args.repeat):
+        idx = torch.randperm(n).cuda()
+        combo_label = torch.zeros(n, num_class).cuda()
+        text_label = torch.zeros(n, num_class).cuda()
+        w = text_classifier.clone()
+        rho = torch.zeros(num_class).cuda()
+        for i in range(n):
+            lr = args.cw / math.sqrt(i + 1)
+            rlr = args.cr / math.sqrt(i + 1)
+            beta = args.beta * math.sqrt((i + 1) / n)
+            x = image_feat[idx[i], :]
+            tlabel = F.softmax(x @ text_classifier / args.tau_t, dim=0)
+            tlabel = tlabel * torch.exp(rho)
+            tlabel /= torch.sum(tlabel)
+            rho -= rlr * (tlabel - args.alpha / num_class)
+            rho[rho < 0] = 0
+            text_label[i, :] = tlabel
+
+            # FIXME margin softmax
+            logits = x @ w / args.tau_i
+            pseudo_y = torch.argmax(tlabel).item()
+            # margin = args.margin if hasattr(args, 'margin') else 0.2
+            logits[pseudo_y] -= margin
+            vision_label = F.softmax(logits, dim=0)
+
+            # vision_label = F.softmax(x @ w / args.tau_i, dim=0)
+            combo_label[i, :] = beta * vision_label + (1 - beta) * tlabel
+            grad = torch.outer(x, vision_label - tlabel)
+            w -= (lr / args.tau_i) * grad
+            w = F.normalize(w, dim=0)
+        acc1, acc5 = accuracy(text_label, image_label[idx], topk=(1, 5))
+        acc_onlab[iter] = (acc1 / n) * 100
+        acc1, acc5 = accuracy(combo_label, image_label[idx], topk=(1, 5))
+
+        # MAPLS - EM Algorithm
+        pz = np.full(len(imagenet_classes), 1.0 / len(imagenet_classes))
+        qy = mapls(combo_label, pz = pz, qy_mode = "soft", max_iter = 100, lam = args.lam)  # FIXME why return nan
+
+        w = np.array(qy) / np.array(pz)
+        if combo_label.is_cuda:
+            combo_label_cpu = combo_label.cpu()
+        qy_probs = lsc(combo_label_cpu, 1.0/w)
+        acc1_ls, acc5_ls = accuracy(qy_probs, image_label[idx], topk=(1, 5))
+
+        acc_onzeta[iter] = (acc1 / n) * 100
+        acc_ls[iter] = (acc1_ls / n) * 100
+    print('mean acc of onlab is: {:.2f}'.format(torch.mean(acc_onlab)))
+    print('mean acc of onzeta is: {:.2f}'.format(torch.mean(acc_onzeta)))
+    print('mean acc of MAPLS is: {:.2f}'.format(torch.mean(acc_ls)))
+
+
+def zeroshot_classifier(clip, model, classnames, templates):
+    with torch.no_grad():
+        zeroshot_weights = []
+        for classname in classnames:
+            texts = [template.format(classname) for template in templates]
+            texts = clip.tokenize(texts).cuda()
+            class_embeddings = model.encode_text(texts)
+            class_embeddings /= class_embeddings.norm(dim=-1, keepdim=True)
+            class_embedding = class_embeddings.mean(dim=0)
+            class_embedding /= class_embedding.norm()
+            zeroshot_weights.append(class_embedding)
+        zeroshot_weights = torch.stack(zeroshot_weights, dim=1).cuda()
+    return zeroshot_weights
+
+
+def accuracy(output, target, topk=(1,)):
+    pred = output.topk(max(topk), 1, True, True)[1].t()
+    pred, target = pred.cpu(), target.cpu()
+    correct = pred.eq(target.view(1, -1).expand_as(pred))
+    return [float(correct[:k].reshape(-1).float().sum(0, keepdim=True).cpu().numpy()) for k in topk]
+
+
+if __name__ == '__main__':
+    # main()
+
+    margins = [1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1, 0.01, 0.001, 0.0001, 0]
+    # margins = [0.01, 0.001, 0.0001]
+    for margin in margins:
+        main(margin)
+