VoiceFilter (code, models, paper)

.gitattributes

@@ -33,3 +33,10 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+Closing[[:space:]]the[[:space:]]Gap[[:space:]]between[[:space:]]Single-User[[:space:]]and[[:space:]]Multi-User[[:space:]]VoiceFilter-Lite.pdf filter=lfs diff=lfs merge=lfs -text
+models/ailia-models/code/output.wav filter=lfs diff=lfs merge=lfs -text
+models/ailia-models/code/ref-voice.wav filter=lfs diff=lfs merge=lfs -text
+Multi-user[[:space:]]VoiceFilter-Lite[[:space:]]via[[:space:]]Attentive[[:space:]]Speaker[[:space:]]Embedding.pdf filter=lfs diff=lfs merge=lfs -text
+Personalized[[:space:]]Keyphrase[[:space:]]Detection[[:space:]]using[[:space:]]Speaker[[:space:]]and[[:space:]]Environment[[:space:]]Information.pdf filter=lfs diff=lfs merge=lfs -text
+VoiceFilter-Lite.[[:space:]]Streaming[[:space:]]Targeted[[:space:]]Voice[[:space:]]Separation[[:space:]]for[[:space:]]On-Device[[:space:]]Speech[[:space:]]Recognition.pdf filter=lfs diff=lfs merge=lfs -text
+VoiceFilter.[[:space:]]Targeted[[:space:]]Voice[[:space:]]Separation[[:space:]]by[[:space:]]Speaker-Conditioned[[:space:]]Spectrogram[[:space:]]Masking.pdf filter=lfs diff=lfs merge=lfs -text

Closing the Gap between Single-User and Multi-User VoiceFilter-Lite.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a10a21a7b1679e924dd38173a202c03c68bcf821fd817708f6445c61c8f8a04
+size 326164

Multi-user VoiceFilter-Lite via Attentive Speaker Embedding.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2323a9abc25e7c32e39daa3a14922191194c2a200f6b1a6ce1b84ce114cddc2f
+size 357161

Personalized Keyphrase Detection using Speaker and Environment Information.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ed2a24c68705c91b106ae5677e6d2bb31015336d6fc18e07d954bed381243ae
+size 202475

VoiceFilter-Lite. Streaming Targeted Voice Separation for On-Device Speech Recognition.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e43753e1b8360fc1d20038e434c7b7d2b24d2e5812e04a6e2e3accd47aa9a4b
+size 397260

VoiceFilter. Targeted Voice Separation by Speaker-Conditioned Spectrogram Masking.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d269f0de0e97c4876adaaa06e9b484362fa99800951cf856fa74e93f70b7799e
+size 259666

code/voicefilter.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6861ee74bafa02fef792d910f69b1eec49333fdf3e1ba4dc7aeb9fb81c51f0a5
+size 2458322

models/ailia-models/code/LICENSE

@@ -0,0 +1,202 @@
+
+                                 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.

models/ailia-models/code/README.md

@@ -0,0 +1,63 @@
+# VoiceFilter
+
+## Input
+
+Audio file
+
+- Mixed audio
+
+https://user-images.githubusercontent.com/29946532/149924386-1a67ec3c-390a-422e-b0b3-9540dff58c72.mov
+
+- Reference audio for d-vector
+
+https://user-images.githubusercontent.com/29946532/149924422-3620f0fb-dca7-45a9-a465-c22bfb00d7ae.mov
+
+Input an audio file that is spoken by multiple people and an audio file that contains the voices of the people you want to extract.
+The voice of one person is extracted and output.
+
+## Output
+
+Audio file
+
+- Estimated audio
+
+https://user-images.githubusercontent.com/29946532/149924467-124a6605-1a52-41ce-8d79-bc54335a0f28.mov
+
+- Ground truth
+
+https://user-images.githubusercontent.com/29946532/149924495-1398c57b-5e8a-4012-8e9a-97e655a3ea26.mov
+
+(Audio from http://swpark.me/voicefilter/)
+
+## Usage
+Automatically downloads the onnx and prototxt files on the first run.
+It is necessary to be connected to the Internet while downloading.
+
+For the sample wav,
+```bash
+$ python3 voicefilter.py --input mixed.wav --reference_file ref-voice.wav
+```
+
+If you want to specify the mixed audio, put the file path after the `--input` option, and to specify the reference audio, put the file path after the `--reference_file` option.   
+You can use `--savepath` option to change the name of the output file to save.
+```bash
+$ python3 voicefilter.py --input MIXED_WAV --reference_file REFERENCE_WAV --savepath SAVE_PATH
+```
+
+
+## Reference
+
+- [VoiceFilter](https://github.com/mindslab-ai/voicefilter)
+
+## Framework
+
+Pytorch
+
+## Model Format
+
+ONNX opset=11
+
+## Netron
+
+[embedder.onnx.prototxt](https://netron.app/?url=https://storage.googleapis.com/ailia-models/voicefilter/embedder.onnx.prototxt)  
+[model.onnx.prototxt](https://netron.app/?url=https://storage.googleapis.com/ailia-models/voicefilter/model.onnx.prototxt)

models/ailia-models/code/audio_utils.py

@@ -0,0 +1,103 @@
+# adapted from Keith Ito's tacotron implementation (MIT License)
+# https://github.com/keithito/tacotron/blob/master/util/audio.py
+
+import librosa
+import numpy as np
+
+
+class Dotdict(dict):
+    """
+    a dictionary that supports dot notation
+    as well as dictionary access notation
+    usage: d = DotDict() or d = DotDict({'val1':'first'})
+    set attributes: d.val2 = 'second' or d['val2'] = 'second'
+    get attributes: d.val2 or d['val2']
+    """
+    __getattr__ = dict.__getitem__
+    __setattr__ = dict.__setitem__
+    __delattr__ = dict.__delitem__
+
+    def __init__(self, dct=None):
+        dct = dict() if not dct else dct
+        for key, value in dct.items():
+            if hasattr(value, 'keys'):
+                value = Dotdict(value)
+            self[key] = value
+
+
+hp = Dotdict({
+    "audio": {
+        "n_fft": 1200,
+        "num_freq": 601,  # n_fft//2 + 1
+        "sample_rate": 16000,
+        "hop_length": 160,
+        "win_length": 400,
+        "min_level_db": -100.0,
+        "ref_level_db": 20.0,
+    },
+    "embedder": {
+        "n_fft": 512,
+        "num_mels": 40,
+    }
+})
+
+
+class Audio:
+    def __init__(self):
+        self.mel_basis = librosa.filters.mel(
+            sr=hp.audio.sample_rate,
+            n_fft=hp.embedder.n_fft,
+            n_mels=hp.embedder.num_mels)
+
+    def get_mel(self, y):
+        y = librosa.core.stft(
+            y=y, n_fft=hp.embedder.n_fft,
+            hop_length=hp.audio.hop_length,
+            win_length=hp.audio.win_length,
+            window='hann')
+        magnitudes = np.abs(y) ** 2
+        mel = np.log10(np.dot(self.mel_basis, magnitudes) + 1e-6)
+
+        return mel
+
+    def wav2spec(self, y):
+        D = self.stft(y)
+        S = self.amp_to_db(np.abs(D)) - hp.audio.ref_level_db
+        S, D = self.normalize(S), np.angle(D)
+        S, D = S.T, D.T  # to make [time, freq]
+
+        return S, D
+
+    def spec2wav(self, spectrogram, phase):
+        spectrogram, phase = spectrogram.T, phase.T
+        # used during inference only
+        # spectrogram: enhanced output
+        # phase: use noisy input's phase, so no GLA is required
+        S = self.db_to_amp(self.denormalize(spectrogram) + hp.audio.ref_level_db)
+
+        return self.istft(S, phase)
+
+    def stft(self, y):
+        return librosa.stft(
+            y=y, n_fft=hp.audio.n_fft,
+            hop_length=hp.audio.hop_length,
+            win_length=hp.audio.win_length)
+
+    def istft(self, mag, phase):
+        stft_matrix = mag * np.exp(1j * phase)
+        return librosa.istft(
+            stft_matrix,
+            hop_length=hp.audio.hop_length,
+            win_length=hp.audio.win_length)
+
+    def amp_to_db(self, x):
+        return 20.0 * np.log10(np.maximum(1e-5, x))
+
+    def db_to_amp(self, x):
+        return np.power(10.0, x * 0.05)
+
+    def normalize(self, S):
+        return np.clip(S / -hp.audio.min_level_db, -1.0, 0.0) + 1.0
+
+    def denormalize(self, S):
+        return (np.clip(S, 0.0, 1.0) - 1.0) * -hp.audio.min_level_db

models/ailia-models/code/output.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f95bc1236025be383e5e3c5406e7f074a9c4176336bfd0f7061b16778696cd22
+size 144044

models/ailia-models/code/ref-voice.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75159cfeba97900bfe9bedbb85a6e89e70d2b5d7e731deae43a4927938a2ad83
+size 148878

models/ailia-models/code/voicefilter.py

@@ -0,0 +1,141 @@
+import os
+import sys
+import time
+
+import numpy as np
+import librosa
+import soundfile as sf
+
+import ailia
+from audio_utils import Audio
+
+# import original modules
+sys.path.append('../../util')
+from arg_utils import get_base_parser, update_parser, get_savepath  # noqa: E402
+from model_utils import check_and_download_models  # noqa: E402
+# logger
+from logging import getLogger  # noqa: E402
+
+logger = getLogger(__name__)
+
+# ======================
+# Parameters
+# ======================
+
+WEIGHT_PATH = 'model.onnx'
+MODEL_PATH = 'model.onnx.prototxt'
+WEIGHT_EMB_PATH = 'embedder.onnx'
+MODEL_EMB_PATH = 'embedder.onnx.prototxt'
+REMOTE_PATH = 'https://storage.googleapis.com/ailia-models/voicefilter/'
+
+WAVE_PATH = "mixed.wav"
+SAVE_PATH = 'output.wav'
+
+# Audio
+SAMPLING_RATE = 16000
+
+# ======================
+# Arguemnt Parser Config
+# ======================
+
+parser = get_base_parser(
+    'VoiceFilter', WAVE_PATH, SAVE_PATH, input_ftype='audio'
+)
+parser.add_argument(
+    '-r', '--reference_file',
+    default="ref-voice.wav", type=str,
+    help='path of reference wav file'
+)
+args = update_parser(parser)
+
+
+# ======================
+# Secondaty Functions
+# ======================
+
+def read_wave(path):
+    # prepare input data
+    wav, source_sr = librosa.load(path, sr=None)
+    # Resample the wav if needed
+    if source_sr is not None and source_sr != SAMPLING_RATE:
+        wav = librosa.resample(wav, orig_sr=source_sr, target_sr=SAMPLING_RATE)
+
+    return wav
+
+
+# ======================
+# Main functions
+# ======================
+
+def audio_recognition(net, embedder):
+    reference_file = args.reference_file
+    if not reference_file or not os.path.exists(reference_file):
+        logger.error('reference_file:%s is NG.' % reference_file)
+        sys.exit(-1)
+
+    audio = Audio()
+
+    # prepare reference wav
+    dvec_wav = read_wave(reference_file)
+    dvec_mel = audio.get_mel(dvec_wav)
+    output = embedder.predict([dvec_mel])
+    dvec = output[0]
+    dvec = np.expand_dims(dvec, axis=0)
+
+    for soundf_path in args.input:
+        logger.info(soundf_path)
+
+        # prepare mix wav
+        mixed_wav = read_wave(soundf_path)
+        mag, phase = audio.wav2spec(mixed_wav)
+        mag = np.expand_dims(mag, axis=0)
+
+        # inference
+        logger.info('Start inference...')
+        if args.benchmark:
+            logger.info('BENCHMARK mode')
+            total_time_estimation = 0
+            for i in range(args.benchmark_count):
+                start = int(round(time.time() * 1000))
+                output = net.predict([mag, dvec])
+                end = int(round(time.time() * 1000))
+                estimation_time = (end - start)
+
+                # Loggin
+                logger.info(f'\tailia processing estimation time {estimation_time} ms')
+                if i != 0:
+                    total_time_estimation = total_time_estimation + estimation_time
+
+            logger.info(f'\taverage time estimation {total_time_estimation / (args.benchmark_count - 1)} ms')
+        else:
+            output = net.predict([mag, dvec])
+
+        mask = output[0]
+
+        est_mag = mag * mask
+        est_wav = audio.spec2wav(est_mag[0], phase)
+
+        savepath = get_savepath(args.savepath, soundf_path, ext='.wav')
+        logger.info(f'saved at : {savepath}')
+        sf.write(savepath, est_wav, SAMPLING_RATE, 'PCM_24')
+
+    logger.info('Script finished successfully.')
+
+
+def main():
+    # model files check and download
+    logger.info('Checking voicefilter model...')
+    check_and_download_models(WEIGHT_PATH, MODEL_PATH, REMOTE_PATH)
+    logger.info('Checking embedder model...')
+    check_and_download_models(WEIGHT_EMB_PATH, MODEL_EMB_PATH, REMOTE_PATH)
+
+    env_id = args.env_id
+
+    net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=env_id)
+    embedder = ailia.Net(MODEL_EMB_PATH, WEIGHT_EMB_PATH, env_id=env_id)
+
+    audio_recognition(net, embedder)
+
+
+if __name__ == '__main__':
+    main()

models/ailia-models/embedder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4fb342dcbd8b8dd9977816343c56d5641a3be86233661e07e4beab79024df872
+size 48654336

models/ailia-models/embedder.onnx.prototxt

@@ -0,0 +1,1179 @@
+ir_version: 7
+producer_name: "pytorch"
+producer_version: "1.10"
+model_version: 0
+graph {
+  name: "torch-jit-export"
+  node {
+    output: "15"
+    name: "Constant_0"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "16"
+    name: "Constant_1"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "17"
+    name: "Constant_2"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "dvec_mel"
+    input: "16"
+    input: "17"
+    input: "15"
+    output: "18"
+    name: "Slice_3"
+    op_type: "Slice"
+  }
+  node {
+    output: "19"
+    name: "Constant_4"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "20"
+    name: "Constant_5"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "21"
+    name: "Constant_6"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "dvec_mel"
+    input: "20"
+    input: "21"
+    input: "19"
+    output: "22"
+    name: "Slice_7"
+    op_type: "Slice"
+  }
+  node {
+    output: "23"
+    name: "Constant_8"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "24"
+    name: "Constant_9"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "25"
+    name: "Constant_10"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "dvec_mel"
+    input: "24"
+    input: "25"
+    input: "23"
+    output: "26"
+    name: "Slice_11"
+    op_type: "Slice"
+  }
+  node {
+    output: "27"
+    name: "Constant_12"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "28"
+    name: "Constant_13"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "29"
+    name: "Constant_14"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "dvec_mel"
+    input: "28"
+    input: "29"
+    input: "27"
+    output: "30"
+    name: "Slice_15"
+    op_type: "Slice"
+  }
+  node {
+    output: "31"
+    name: "Constant_16"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "32"
+    name: "Constant_17"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "33"
+    name: "Constant_18"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "dvec_mel"
+    input: "32"
+    input: "33"
+    input: "31"
+    output: "34"
+    name: "Slice_19"
+    op_type: "Slice"
+  }
+  node {
+    output: "35"
+    name: "Constant_20"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "36"
+    name: "Constant_21"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "37"
+    name: "Constant_22"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "dvec_mel"
+    input: "36"
+    input: "37"
+    input: "35"
+    output: "38"
+    name: "Slice_23"
+    op_type: "Slice"
+  }
+  node {
+    input: "18"
+    output: "39"
+    name: "Unsqueeze_24"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "22"
+    output: "40"
+    name: "Unsqueeze_25"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "26"
+    output: "41"
+    name: "Unsqueeze_26"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "30"
+    output: "42"
+    name: "Unsqueeze_27"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "34"
+    output: "43"
+    name: "Unsqueeze_28"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "38"
+    output: "44"
+    name: "Unsqueeze_29"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "39"
+    input: "40"
+    input: "41"
+    input: "42"
+    input: "43"
+    input: "44"
+    output: "45"
+    name: "Concat_30"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 1
+      type: INT
+    }
+  }
+  node {
+    input: "45"
+    output: "48"
+    name: "Transpose_31"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 2
+      ints: 1
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    input: "48"
+    output: "114"
+    name: "Shape_32"
+    op_type: "Shape"
+  }
+  node {
+    output: "115"
+    name: "Constant_33"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "114"
+    input: "115"
+    output: "116"
+    name: "Gather_34"
+    op_type: "Gather"
+  }
+  node {
+    input: "116"
+    output: "117"
+    name: "Unsqueeze_35"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "118"
+    name: "Constant_36"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "340"
+    input: "117"
+    input: "118"
+    output: "121"
+    name: "Concat_37"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "338"
+    input: "121"
+    output: "122"
+    name: "Expand_38"
+    op_type: "Expand"
+  }
+  node {
+    input: "48"
+    output: "123"
+    name: "Shape_39"
+    op_type: "Shape"
+  }
+  node {
+    output: "124"
+    name: "Constant_40"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "123"
+    input: "124"
+    output: "125"
+    name: "Gather_41"
+    op_type: "Gather"
+  }
+  node {
+    input: "125"
+    output: "126"
+    name: "Unsqueeze_42"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "127"
+    name: "Constant_43"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "341"
+    input: "126"
+    input: "127"
+    output: "130"
+    name: "Concat_44"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "339"
+    input: "130"
+    output: "131"
+    name: "Expand_45"
+    op_type: "Expand"
+  }
+  node {
+    input: "48"
+    input: "335"
+    input: "336"
+    input: "337"
+    input: ""
+    input: "122"
+    input: "131"
+    output: "132"
+    output: "133"
+    output: "134"
+    name: "LSTM_46"
+    op_type: "LSTM"
+    attribute {
+      name: "hidden_size"
+      i: 768
+      type: INT
+    }
+  }
+  node {
+    input: "132"
+    output: "135"
+    name: "Squeeze_47"
+    op_type: "Squeeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "135"
+    output: "200"
+    name: "Shape_48"
+    op_type: "Shape"
+  }
+  node {
+    output: "201"
+    name: "Constant_49"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "200"
+    input: "201"
+    output: "202"
+    name: "Gather_50"
+    op_type: "Gather"
+  }
+  node {
+    input: "202"
+    output: "203"
+    name: "Unsqueeze_51"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "204"
+    name: "Constant_52"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "364"
+    input: "203"
+    input: "204"
+    output: "207"
+    name: "Concat_53"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "362"
+    input: "207"
+    output: "208"
+    name: "Expand_54"
+    op_type: "Expand"
+  }
+  node {
+    input: "135"
+    output: "209"
+    name: "Shape_55"
+    op_type: "Shape"
+  }
+  node {
+    output: "210"
+    name: "Constant_56"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "209"
+    input: "210"
+    output: "211"
+    name: "Gather_57"
+    op_type: "Gather"
+  }
+  node {
+    input: "211"
+    output: "212"
+    name: "Unsqueeze_58"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "213"
+    name: "Constant_59"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "365"
+    input: "212"
+    input: "213"
+    output: "216"
+    name: "Concat_60"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "363"
+    input: "216"
+    output: "217"
+    name: "Expand_61"
+    op_type: "Expand"
+  }
+  node {
+    input: "135"
+    input: "359"
+    input: "360"
+    input: "361"
+    input: ""
+    input: "208"
+    input: "217"
+    output: "218"
+    output: "219"
+    output: "220"
+    name: "LSTM_62"
+    op_type: "LSTM"
+    attribute {
+      name: "hidden_size"
+      i: 768
+      type: INT
+    }
+  }
+  node {
+    input: "218"
+    output: "221"
+    name: "Squeeze_63"
+    op_type: "Squeeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "221"
+    output: "286"
+    name: "Shape_64"
+    op_type: "Shape"
+  }
+  node {
+    output: "287"
+    name: "Constant_65"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "286"
+    input: "287"
+    output: "288"
+    name: "Gather_66"
+    op_type: "Gather"
+  }
+  node {
+    input: "288"
+    output: "289"
+    name: "Unsqueeze_67"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "290"
+    name: "Constant_68"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "388"
+    input: "289"
+    input: "290"
+    output: "293"
+    name: "Concat_69"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "386"
+    input: "293"
+    output: "294"
+    name: "Expand_70"
+    op_type: "Expand"
+  }
+  node {
+    input: "221"
+    output: "295"
+    name: "Shape_71"
+    op_type: "Shape"
+  }
+  node {
+    output: "296"
+    name: "Constant_72"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "295"
+    input: "296"
+    output: "297"
+    name: "Gather_73"
+    op_type: "Gather"
+  }
+  node {
+    input: "297"
+    output: "298"
+    name: "Unsqueeze_74"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "299"
+    name: "Constant_75"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "389"
+    input: "298"
+    input: "299"
+    output: "302"
+    name: "Concat_76"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "387"
+    input: "302"
+    output: "303"
+    name: "Expand_77"
+    op_type: "Expand"
+  }
+  node {
+    input: "221"
+    input: "383"
+    input: "384"
+    input: "385"
+    input: ""
+    input: "294"
+    input: "303"
+    output: "304"
+    output: "305"
+    output: "306"
+    name: "LSTM_78"
+    op_type: "LSTM"
+    attribute {
+      name: "hidden_size"
+      i: 768
+      type: INT
+    }
+  }
+  node {
+    input: "304"
+    output: "307"
+    name: "Squeeze_79"
+    op_type: "Squeeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "307"
+    output: "308"
+    name: "Transpose_80"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      ints: 2
+      type: INTS
+    }
+  }
+  node {
+    output: "309"
+    name: "Constant_81"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "308"
+    input: "309"
+    output: "310"
+    name: "Gather_82"
+    op_type: "Gather"
+    attribute {
+      name: "axis"
+      i: 1
+      type: INT
+    }
+  }
+  node {
+    input: "310"
+    input: "proj.linear_layer.weight"
+    input: "proj.linear_layer.bias"
+    output: "311"
+    name: "Gemm_83"
+    op_type: "Gemm"
+    attribute {
+      name: "alpha"
+      f: 1.0
+      type: FLOAT
+    }
+    attribute {
+      name: "beta"
+      f: 1.0
+      type: FLOAT
+    }
+    attribute {
+      name: "transB"
+      i: 1
+      type: INT
+    }
+  }
+  node {
+    input: "311"
+    output: "312"
+    name: "ReduceL2_84"
+    op_type: "ReduceL2"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "keepdims"
+      i: 1
+      type: INT
+    }
+  }
+  node {
+    input: "311"
+    input: "312"
+    output: "313"
+    name: "Div_85"
+    op_type: "Div"
+  }
+  node {
+    input: "313"
+    output: "314"
+    name: "ReduceSum_86"
+    op_type: "ReduceSum"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "keepdims"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "314"
+    input: "390"
+    output: "dvec"
+    name: "Div_87"
+    op_type: "Div"
+  }
+  initializer {
+      dims: 256
+      dims: 768
+      data_type: 1
+      name: "proj.linear_layer.weight"
+  }
+  initializer {
+      dims: 256
+      data_type: 1
+      name: "proj.linear_layer.bias"
+  }
+  initializer {
+      dims: 1
+      dims: 3072
+      dims: 40
+      data_type: 1
+      name: "335"
+  }
+  initializer {
+      dims: 1
+      dims: 3072
+      dims: 768
+      data_type: 1
+      name: "336"
+  }
+  initializer {
+      dims: 1
+      dims: 6144
+      data_type: 1
+      name: "337"
+  }
+  initializer {
+      dims: 1
+      dims: 6
+      dims: 768
+      data_type: 1
+      name: "338"
+  }
+  initializer {
+      dims: 1
+      dims: 6
+      dims: 768
+      data_type: 1
+      name: "339"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "340"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "341"
+  }
+  initializer {
+      dims: 1
+      dims: 3072
+      dims: 768
+      data_type: 1
+      name: "359"
+  }
+  initializer {
+      dims: 1
+      dims: 3072
+      dims: 768
+      data_type: 1
+      name: "360"
+  }
+  initializer {
+      dims: 1
+      dims: 6144
+      data_type: 1
+      name: "361"
+  }
+  initializer {
+      dims: 1
+      dims: 6
+      dims: 768
+      data_type: 1
+      name: "362"
+  }
+  initializer {
+      dims: 1
+      dims: 6
+      dims: 768
+      data_type: 1
+      name: "363"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "364"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "365"
+  }
+  initializer {
+      dims: 1
+      dims: 3072
+      dims: 768
+      data_type: 1
+      name: "383"
+  }
+  initializer {
+      dims: 1
+      dims: 3072
+      dims: 768
+      data_type: 1
+      name: "384"
+  }
+  initializer {
+      dims: 1
+      dims: 6144
+      data_type: 1
+      name: "385"
+  }
+  initializer {
+      dims: 1
+      dims: 6
+      dims: 768
+      data_type: 1
+      name: "386"
+  }
+  initializer {
+      dims: 1
+      dims: 6
+      dims: 768
+      data_type: 1
+      name: "387"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "388"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "389"
+  }
+  initializer {
+      data_type: 1
+      name: "390"
+  }
+  input {
+    name: "dvec_mel"
+    type {
+      tensor_type {
+        elem_type: 1
+        shape {
+          dim {
+            dim_value: 40
+          }
+          dim {
+            dim_value: 301
+          }
+        }
+      }
+    }
+  }
+  output {
+    name: "dvec"
+    type {
+      tensor_type {
+        elem_type: 1
+        shape {
+          dim {
+            dim_value: 256
+          }
+        }
+      }
+    }
+  }
+}
+opset_import {
+  version: 11
+}

models/ailia-models/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:698223eb28f14536c0a20b3b1169470fc1331f393a8dcf31a4205e78a7acfe4e
+size 75510811

models/ailia-models/model.onnx.prototxt

@@ -0,0 +1,2425 @@
+ir_version: 7
+producer_name: "pytorch"
+producer_version: "1.10"
+model_version: 0
+graph {
+  name: "torch-jit-export"
+  node {
+    input: "mag"
+    output: "70"
+    name: "Unsqueeze_0"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "472"
+    output: "80"
+    name: "ConstantOfShape_1"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+        raw_data: "\000\000\000\000\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "473"
+    input: "80"
+    output: "81"
+    name: "Concat_2"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    output: "82"
+    name: "Constant_3"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 2
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "81"
+    input: "82"
+    output: "83"
+    name: "Reshape_4"
+    op_type: "Reshape"
+  }
+  node {
+    output: "84"
+    name: "Constant_5"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "85"
+    name: "Constant_6"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "86"
+    name: "Constant_7"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "87"
+    name: "Constant_8"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "83"
+    input: "85"
+    input: "86"
+    input: "84"
+    input: "87"
+    output: "88"
+    name: "Slice_9"
+    op_type: "Slice"
+  }
+  node {
+    input: "88"
+    output: "89"
+    name: "Transpose_10"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "90"
+    name: "Constant_11"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "89"
+    input: "90"
+    output: "91"
+    name: "Reshape_12"
+    op_type: "Reshape"
+  }
+  node {
+    input: "91"
+    output: "92"
+    name: "Cast_13"
+    op_type: "Cast"
+    attribute {
+      name: "to"
+      i: 7
+      type: INT
+    }
+  }
+  node {
+    output: "93"
+    name: "Constant_14"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 1
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "70"
+    input: "92"
+    input: "93"
+    output: "94"
+    name: "Pad_15"
+    op_type: "Pad"
+    attribute {
+      name: "mode"
+      s: "constant"
+      type: STRING
+    }
+  }
+  node {
+    input: "94"
+    input: "446"
+    input: "447"
+    output: "445"
+    name: "Conv_16"
+    op_type: "Conv"
+    attribute {
+      name: "dilations"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "group"
+      i: 1
+      type: INT
+    }
+    attribute {
+      name: "kernel_shape"
+      ints: 1
+      ints: 7
+      type: INTS
+    }
+    attribute {
+      name: "pads"
+      ints: 0
+      ints: 0
+      ints: 0
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "strides"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "445"
+    output: "97"
+    name: "Relu_17"
+    op_type: "Relu"
+  }
+  node {
+    input: "477"
+    output: "107"
+    name: "ConstantOfShape_18"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+        raw_data: "\000\000\000\000\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "478"
+    input: "107"
+    output: "108"
+    name: "Concat_19"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    output: "109"
+    name: "Constant_20"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 2
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "108"
+    input: "109"
+    output: "110"
+    name: "Reshape_21"
+    op_type: "Reshape"
+  }
+  node {
+    output: "111"
+    name: "Constant_22"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "112"
+    name: "Constant_23"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "113"
+    name: "Constant_24"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "114"
+    name: "Constant_25"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "110"
+    input: "112"
+    input: "113"
+    input: "111"
+    input: "114"
+    output: "115"
+    name: "Slice_26"
+    op_type: "Slice"
+  }
+  node {
+    input: "115"
+    output: "116"
+    name: "Transpose_27"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "117"
+    name: "Constant_28"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "116"
+    input: "117"
+    output: "118"
+    name: "Reshape_29"
+    op_type: "Reshape"
+  }
+  node {
+    input: "118"
+    output: "119"
+    name: "Cast_30"
+    op_type: "Cast"
+    attribute {
+      name: "to"
+      i: 7
+      type: INT
+    }
+  }
+  node {
+    output: "120"
+    name: "Constant_31"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 1
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "97"
+    input: "119"
+    input: "120"
+    output: "121"
+    name: "Pad_32"
+    op_type: "Pad"
+    attribute {
+      name: "mode"
+      s: "constant"
+      type: STRING
+    }
+  }
+  node {
+    input: "121"
+    input: "449"
+    input: "450"
+    output: "448"
+    name: "Conv_33"
+    op_type: "Conv"
+    attribute {
+      name: "dilations"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "group"
+      i: 1
+      type: INT
+    }
+    attribute {
+      name: "kernel_shape"
+      ints: 7
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "pads"
+      ints: 0
+      ints: 0
+      ints: 0
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "strides"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "448"
+    output: "124"
+    name: "Relu_34"
+    op_type: "Relu"
+  }
+  node {
+    input: "482"
+    output: "134"
+    name: "ConstantOfShape_35"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+        raw_data: "\000\000\000\000\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "483"
+    input: "134"
+    output: "135"
+    name: "Concat_36"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    output: "136"
+    name: "Constant_37"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 2
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "135"
+    input: "136"
+    output: "137"
+    name: "Reshape_38"
+    op_type: "Reshape"
+  }
+  node {
+    output: "138"
+    name: "Constant_39"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "139"
+    name: "Constant_40"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "140"
+    name: "Constant_41"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "141"
+    name: "Constant_42"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "137"
+    input: "139"
+    input: "140"
+    input: "138"
+    input: "141"
+    output: "142"
+    name: "Slice_43"
+    op_type: "Slice"
+  }
+  node {
+    input: "142"
+    output: "143"
+    name: "Transpose_44"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "144"
+    name: "Constant_45"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "143"
+    input: "144"
+    output: "145"
+    name: "Reshape_46"
+    op_type: "Reshape"
+  }
+  node {
+    input: "145"
+    output: "146"
+    name: "Cast_47"
+    op_type: "Cast"
+    attribute {
+      name: "to"
+      i: 7
+      type: INT
+    }
+  }
+  node {
+    output: "147"
+    name: "Constant_48"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 1
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "124"
+    input: "146"
+    input: "147"
+    output: "148"
+    name: "Pad_49"
+    op_type: "Pad"
+    attribute {
+      name: "mode"
+      s: "constant"
+      type: STRING
+    }
+  }
+  node {
+    input: "148"
+    input: "452"
+    input: "453"
+    output: "451"
+    name: "Conv_50"
+    op_type: "Conv"
+    attribute {
+      name: "dilations"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "group"
+      i: 1
+      type: INT
+    }
+    attribute {
+      name: "kernel_shape"
+      ints: 5
+      ints: 5
+      type: INTS
+    }
+    attribute {
+      name: "pads"
+      ints: 0
+      ints: 0
+      ints: 0
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "strides"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "451"
+    output: "151"
+    name: "Relu_51"
+    op_type: "Relu"
+  }
+  node {
+    input: "487"
+    output: "161"
+    name: "ConstantOfShape_52"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+        raw_data: "\000\000\000\000\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "488"
+    input: "161"
+    output: "162"
+    name: "Concat_53"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    output: "163"
+    name: "Constant_54"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 2
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "162"
+    input: "163"
+    output: "164"
+    name: "Reshape_55"
+    op_type: "Reshape"
+  }
+  node {
+    output: "165"
+    name: "Constant_56"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "166"
+    name: "Constant_57"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "167"
+    name: "Constant_58"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "168"
+    name: "Constant_59"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "164"
+    input: "166"
+    input: "167"
+    input: "165"
+    input: "168"
+    output: "169"
+    name: "Slice_60"
+    op_type: "Slice"
+  }
+  node {
+    input: "169"
+    output: "170"
+    name: "Transpose_61"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "171"
+    name: "Constant_62"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "170"
+    input: "171"
+    output: "172"
+    name: "Reshape_63"
+    op_type: "Reshape"
+  }
+  node {
+    input: "172"
+    output: "173"
+    name: "Cast_64"
+    op_type: "Cast"
+    attribute {
+      name: "to"
+      i: 7
+      type: INT
+    }
+  }
+  node {
+    output: "174"
+    name: "Constant_65"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 1
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "151"
+    input: "173"
+    input: "174"
+    output: "175"
+    name: "Pad_66"
+    op_type: "Pad"
+    attribute {
+      name: "mode"
+      s: "constant"
+      type: STRING
+    }
+  }
+  node {
+    input: "175"
+    input: "455"
+    input: "456"
+    output: "454"
+    name: "Conv_67"
+    op_type: "Conv"
+    attribute {
+      name: "dilations"
+      ints: 2
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "group"
+      i: 1
+      type: INT
+    }
+    attribute {
+      name: "kernel_shape"
+      ints: 5
+      ints: 5
+      type: INTS
+    }
+    attribute {
+      name: "pads"
+      ints: 0
+      ints: 0
+      ints: 0
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "strides"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "454"
+    output: "178"
+    name: "Relu_68"
+    op_type: "Relu"
+  }
+  node {
+    input: "492"
+    output: "188"
+    name: "ConstantOfShape_69"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+        raw_data: "\000\000\000\000\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "493"
+    input: "188"
+    output: "189"
+    name: "Concat_70"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    output: "190"
+    name: "Constant_71"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 2
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "189"
+    input: "190"
+    output: "191"
+    name: "Reshape_72"
+    op_type: "Reshape"
+  }
+  node {
+    output: "192"
+    name: "Constant_73"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "193"
+    name: "Constant_74"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "194"
+    name: "Constant_75"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "195"
+    name: "Constant_76"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "191"
+    input: "193"
+    input: "194"
+    input: "192"
+    input: "195"
+    output: "196"
+    name: "Slice_77"
+    op_type: "Slice"
+  }
+  node {
+    input: "196"
+    output: "197"
+    name: "Transpose_78"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "198"
+    name: "Constant_79"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "197"
+    input: "198"
+    output: "199"
+    name: "Reshape_80"
+    op_type: "Reshape"
+  }
+  node {
+    input: "199"
+    output: "200"
+    name: "Cast_81"
+    op_type: "Cast"
+    attribute {
+      name: "to"
+      i: 7
+      type: INT
+    }
+  }
+  node {
+    output: "201"
+    name: "Constant_82"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 1
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "178"
+    input: "200"
+    input: "201"
+    output: "202"
+    name: "Pad_83"
+    op_type: "Pad"
+    attribute {
+      name: "mode"
+      s: "constant"
+      type: STRING
+    }
+  }
+  node {
+    input: "202"
+    input: "458"
+    input: "459"
+    output: "457"
+    name: "Conv_84"
+    op_type: "Conv"
+    attribute {
+      name: "dilations"
+      ints: 4
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "group"
+      i: 1
+      type: INT
+    }
+    attribute {
+      name: "kernel_shape"
+      ints: 5
+      ints: 5
+      type: INTS
+    }
+    attribute {
+      name: "pads"
+      ints: 0
+      ints: 0
+      ints: 0
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "strides"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "457"
+    output: "205"
+    name: "Relu_85"
+    op_type: "Relu"
+  }
+  node {
+    input: "497"
+    output: "215"
+    name: "ConstantOfShape_86"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+        raw_data: "\000\000\000\000\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "498"
+    input: "215"
+    output: "216"
+    name: "Concat_87"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    output: "217"
+    name: "Constant_88"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 2
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "216"
+    input: "217"
+    output: "218"
+    name: "Reshape_89"
+    op_type: "Reshape"
+  }
+  node {
+    output: "219"
+    name: "Constant_90"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "220"
+    name: "Constant_91"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "221"
+    name: "Constant_92"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "222"
+    name: "Constant_93"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "218"
+    input: "220"
+    input: "221"
+    input: "219"
+    input: "222"
+    output: "223"
+    name: "Slice_94"
+    op_type: "Slice"
+  }
+  node {
+    input: "223"
+    output: "224"
+    name: "Transpose_95"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "225"
+    name: "Constant_96"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "224"
+    input: "225"
+    output: "226"
+    name: "Reshape_97"
+    op_type: "Reshape"
+  }
+  node {
+    input: "226"
+    output: "227"
+    name: "Cast_98"
+    op_type: "Cast"
+    attribute {
+      name: "to"
+      i: 7
+      type: INT
+    }
+  }
+  node {
+    output: "228"
+    name: "Constant_99"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 1
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "205"
+    input: "227"
+    input: "228"
+    output: "229"
+    name: "Pad_100"
+    op_type: "Pad"
+    attribute {
+      name: "mode"
+      s: "constant"
+      type: STRING
+    }
+  }
+  node {
+    input: "229"
+    input: "461"
+    input: "462"
+    output: "460"
+    name: "Conv_101"
+    op_type: "Conv"
+    attribute {
+      name: "dilations"
+      ints: 8
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "group"
+      i: 1
+      type: INT
+    }
+    attribute {
+      name: "kernel_shape"
+      ints: 5
+      ints: 5
+      type: INTS
+    }
+    attribute {
+      name: "pads"
+      ints: 0
+      ints: 0
+      ints: 0
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "strides"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "460"
+    output: "232"
+    name: "Relu_102"
+    op_type: "Relu"
+  }
+  node {
+    input: "502"
+    output: "242"
+    name: "ConstantOfShape_103"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+        raw_data: "\000\000\000\000\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "503"
+    input: "242"
+    output: "243"
+    name: "Concat_104"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    output: "244"
+    name: "Constant_105"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 2
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "243"
+    input: "244"
+    output: "245"
+    name: "Reshape_106"
+    op_type: "Reshape"
+  }
+  node {
+    output: "246"
+    name: "Constant_107"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "247"
+    name: "Constant_108"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "248"
+    name: "Constant_109"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    output: "249"
+    name: "Constant_110"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "245"
+    input: "247"
+    input: "248"
+    input: "246"
+    input: "249"
+    output: "250"
+    name: "Slice_111"
+    op_type: "Slice"
+  }
+  node {
+    input: "250"
+    output: "251"
+    name: "Transpose_112"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    output: "252"
+    name: "Constant_113"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "251"
+    input: "252"
+    output: "253"
+    name: "Reshape_114"
+    op_type: "Reshape"
+  }
+  node {
+    input: "253"
+    output: "254"
+    name: "Cast_115"
+    op_type: "Cast"
+    attribute {
+      name: "to"
+      i: 7
+      type: INT
+    }
+  }
+  node {
+    output: "255"
+    name: "Constant_116"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 1
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "232"
+    input: "254"
+    input: "255"
+    output: "256"
+    name: "Pad_117"
+    op_type: "Pad"
+    attribute {
+      name: "mode"
+      s: "constant"
+      type: STRING
+    }
+  }
+  node {
+    input: "256"
+    input: "464"
+    input: "465"
+    output: "463"
+    name: "Conv_118"
+    op_type: "Conv"
+    attribute {
+      name: "dilations"
+      ints: 16
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "group"
+      i: 1
+      type: INT
+    }
+    attribute {
+      name: "kernel_shape"
+      ints: 5
+      ints: 5
+      type: INTS
+    }
+    attribute {
+      name: "pads"
+      ints: 0
+      ints: 0
+      ints: 0
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "strides"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "463"
+    output: "259"
+    name: "Relu_119"
+    op_type: "Relu"
+  }
+  node {
+    input: "259"
+    input: "467"
+    input: "468"
+    output: "466"
+    name: "Conv_120"
+    op_type: "Conv"
+    attribute {
+      name: "dilations"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "group"
+      i: 1
+      type: INT
+    }
+    attribute {
+      name: "kernel_shape"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+    attribute {
+      name: "pads"
+      ints: 0
+      ints: 0
+      ints: 0
+      ints: 0
+      type: INTS
+    }
+    attribute {
+      name: "strides"
+      ints: 1
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "466"
+    output: "262"
+    name: "Relu_121"
+    op_type: "Relu"
+  }
+  node {
+    input: "262"
+    output: "263"
+    name: "Transpose_122"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 0
+      ints: 2
+      ints: 1
+      ints: 3
+      type: INTS
+    }
+  }
+  node {
+    input: "263"
+    output: "264"
+    name: "Shape_123"
+    op_type: "Shape"
+  }
+  node {
+    output: "265"
+    name: "Constant_124"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "264"
+    input: "265"
+    output: "266"
+    name: "Gather_125"
+    op_type: "Gather"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "263"
+    output: "267"
+    name: "Shape_126"
+    op_type: "Shape"
+  }
+  node {
+    output: "268"
+    name: "Constant_127"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "267"
+    input: "268"
+    output: "269"
+    name: "Gather_128"
+    op_type: "Gather"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "266"
+    output: "271"
+    name: "Unsqueeze_129"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    input: "269"
+    output: "272"
+    name: "Unsqueeze_130"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    input: "271"
+    input: "272"
+    input: "504"
+    output: "274"
+    name: "Concat_131"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "263"
+    input: "274"
+    output: "275"
+    name: "Reshape_132"
+    op_type: "Reshape"
+  }
+  node {
+    input: "dvec"
+    output: "276"
+    name: "Unsqueeze_133"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 1
+      type: INTS
+    }
+  }
+  node {
+    input: "275"
+    output: "277"
+    name: "Shape_134"
+    op_type: "Shape"
+  }
+  node {
+    output: "278"
+    name: "Constant_135"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "277"
+    input: "278"
+    output: "279"
+    name: "Gather_136"
+    op_type: "Gather"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "279"
+    output: "283"
+    name: "Unsqueeze_137"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    input: "505"
+    input: "283"
+    input: "506"
+    output: "285"
+    name: "Concat_138"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "279"
+    output: "287"
+    name: "Unsqueeze_139"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    input: "507"
+    input: "287"
+    input: "508"
+    output: "289"
+    name: "Concat_140"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "285"
+    output: "290"
+    name: "Shape_141"
+    op_type: "Shape"
+  }
+  node {
+    input: "290"
+    output: "291"
+    name: "ConstantOfShape_142"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 7
+        raw_data: "\001\000\000\000\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "276"
+    input: "291"
+    output: "292"
+    name: "Expand_143"
+    op_type: "Expand"
+  }
+  node {
+    input: "292"
+    input: "289"
+    output: "293"
+    name: "Tile_144"
+    op_type: "Tile"
+  }
+  node {
+    input: "275"
+    input: "293"
+    output: "294"
+    name: "Concat_145"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 2
+      type: INT
+    }
+  }
+  node {
+    input: "294"
+    output: "295"
+    name: "Shape_146"
+    op_type: "Shape"
+  }
+  node {
+    output: "296"
+    name: "Constant_147"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "295"
+    input: "296"
+    output: "297"
+    name: "Gather_148"
+    op_type: "Gather"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "297"
+    output: "301"
+    name: "Unsqueeze_149"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    input: "509"
+    input: "301"
+    input: "510"
+    output: "303"
+    name: "Concat_150"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "303"
+    output: "304"
+    name: "ConstantOfShape_151"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 1
+        raw_data: "\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "297"
+    output: "308"
+    name: "Unsqueeze_152"
+    op_type: "Unsqueeze"
+    attribute {
+      name: "axes"
+      ints: 0
+      type: INTS
+    }
+  }
+  node {
+    input: "511"
+    input: "308"
+    input: "512"
+    output: "310"
+    name: "Concat_153"
+    op_type: "Concat"
+    attribute {
+      name: "axis"
+      i: 0
+      type: INT
+    }
+  }
+  node {
+    input: "310"
+    output: "311"
+    name: "ConstantOfShape_154"
+    op_type: "ConstantOfShape"
+    attribute {
+      name: "value"
+      t {
+        dims: 1
+        data_type: 1
+        raw_data: "\000\000\000\000"
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "294"
+    output: "312"
+    name: "Transpose_155"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      ints: 2
+      type: INTS
+    }
+  }
+  node {
+    input: "312"
+    input: "554"
+    input: "555"
+    input: "553"
+    input: ""
+    input: "304"
+    input: "311"
+    output: "429"
+    output: "430"
+    output: "431"
+    name: "LSTM_156"
+    op_type: "LSTM"
+    attribute {
+      name: "direction"
+      s: "bidirectional"
+      type: STRING
+    }
+    attribute {
+      name: "hidden_size"
+      i: 400
+      type: INT
+    }
+  }
+  node {
+    input: "429"
+    output: "432"
+    name: "Transpose_157"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 0
+      ints: 2
+      ints: 1
+      ints: 3
+      type: INTS
+    }
+  }
+  node {
+    output: "433"
+    name: "Constant_158"
+    op_type: "Constant"
+    attribute {
+      name: "value"
+      t {
+        dims: 3
+        data_type: 7
+      }
+      type: TENSOR
+    }
+  }
+  node {
+    input: "432"
+    input: "433"
+    output: "434"
+    name: "Reshape_159"
+    op_type: "Reshape"
+  }
+  node {
+    input: "434"
+    output: "435"
+    name: "Transpose_160"
+    op_type: "Transpose"
+    attribute {
+      name: "perm"
+      ints: 1
+      ints: 0
+      ints: 2
+      type: INTS
+    }
+  }
+  node {
+    input: "435"
+    output: "436"
+    name: "Relu_161"
+    op_type: "Relu"
+  }
+  node {
+    input: "436"
+    input: "556"
+    output: "438"
+    name: "MatMul_162"
+    op_type: "MatMul"
+  }
+  node {
+    input: "fc1.bias"
+    input: "438"
+    output: "439"
+    name: "Add_163"
+    op_type: "Add"
+  }
+  node {
+    input: "439"
+    output: "440"
+    name: "Relu_164"
+    op_type: "Relu"
+  }
+  node {
+    input: "440"
+    input: "557"
+    output: "442"
+    name: "MatMul_165"
+    op_type: "MatMul"
+  }
+  node {
+    input: "fc2.bias"
+    input: "442"
+    output: "443"
+    name: "Add_166"
+    op_type: "Add"
+  }
+  node {
+    input: "443"
+    output: "mask"
+    name: "Sigmoid_167"
+    op_type: "Sigmoid"
+  }
+  initializer {
+      dims: 600
+      data_type: 1
+      name: "fc1.bias"
+  }
+  initializer {
+      dims: 601
+      data_type: 1
+      name: "fc2.bias"
+  }
+  initializer {
+      dims: 64
+      dims: 1
+      dims: 1
+      dims: 7
+      data_type: 1
+      name: "446"
+  }
+  initializer {
+      dims: 64
+      data_type: 1
+      name: "447"
+  }
+  initializer {
+      dims: 64
+      dims: 64
+      dims: 7
+      dims: 1
+      data_type: 1
+      name: "449"
+  }
+  initializer {
+      dims: 64
+      data_type: 1
+      name: "450"
+  }
+  initializer {
+      dims: 64
+      dims: 64
+      dims: 5
+      dims: 5
+      data_type: 1
+      name: "452"
+  }
+  initializer {
+      dims: 64
+      data_type: 1
+      name: "453"
+  }
+  initializer {
+      dims: 64
+      dims: 64
+      dims: 5
+      dims: 5
+      data_type: 1
+      name: "455"
+  }
+  initializer {
+      dims: 64
+      data_type: 1
+      name: "456"
+  }
+  initializer {
+      dims: 64
+      dims: 64
+      dims: 5
+      dims: 5
+      data_type: 1
+      name: "458"
+  }
+  initializer {
+      dims: 64
+      data_type: 1
+      name: "459"
+  }
+  initializer {
+      dims: 64
+      dims: 64
+      dims: 5
+      dims: 5
+      data_type: 1
+      name: "461"
+  }
+  initializer {
+      dims: 64
+      data_type: 1
+      name: "462"
+  }
+  initializer {
+      dims: 64
+      dims: 64
+      dims: 5
+      dims: 5
+      data_type: 1
+      name: "464"
+  }
+  initializer {
+      dims: 64
+      data_type: 1
+      name: "465"
+  }
+  initializer {
+      dims: 8
+      dims: 64
+      dims: 1
+      dims: 1
+      data_type: 1
+      name: "467"
+  }
+  initializer {
+      dims: 8
+      data_type: 1
+      name: "468"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "472"
+  }
+  initializer {
+      dims: 4
+      data_type: 7
+      name: "473"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "477"
+  }
+  initializer {
+      dims: 4
+      data_type: 7
+      name: "478"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "482"
+  }
+  initializer {
+      dims: 4
+      data_type: 7
+      name: "483"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "487"
+  }
+  initializer {
+      dims: 4
+      data_type: 7
+      name: "488"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "492"
+  }
+  initializer {
+      dims: 4
+      data_type: 7
+      name: "493"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "497"
+  }
+  initializer {
+      dims: 4
+      data_type: 7
+      name: "498"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "502"
+  }
+  initializer {
+      dims: 4
+      data_type: 7
+      name: "503"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "504"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "505"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "506"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "507"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "508"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "509"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "510"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "511"
+  }
+  initializer {
+      dims: 1
+      data_type: 7
+      name: "512"
+  }
+  initializer {
+      dims: 2
+      dims: 3200
+      data_type: 1
+      name: "553"
+  }
+  initializer {
+      dims: 2
+      dims: 1600
+      dims: 5064
+      data_type: 1
+      name: "554"
+  }
+  initializer {
+      dims: 2
+      dims: 1600
+      dims: 400
+      data_type: 1
+      name: "555"
+  }
+  initializer {
+      dims: 800
+      dims: 600
+      data_type: 1
+      name: "556"
+  }
+  initializer {
+      dims: 600
+      dims: 601
+      data_type: 1
+      name: "557"
+  }
+  input {
+    name: "mag"
+    type {
+      tensor_type {
+        elem_type: 1
+        shape {
+          dim {
+            dim_value: 1
+          }
+          dim {
+            dim_param: "L"
+          }
+          dim {
+            dim_value: 601
+          }
+        }
+      }
+    }
+  }
+  input {
+    name: "dvec"
+    type {
+      tensor_type {
+        elem_type: 1
+        shape {
+          dim {
+            dim_value: 1
+          }
+          dim {
+            dim_value: 256
+          }
+        }
+      }
+    }
+  }
+  output {
+    name: "mask"
+    type {
+      tensor_type {
+        elem_type: 1
+        shape {
+          dim {
+            dim_param: "1"
+          }
+          dim {
+            dim_param: "L"
+          }
+          dim {
+            dim_value: 601
+          }
+        }
+      }
+    }
+  }
+}
+opset_import {
+  version: 11
+}

models/ailia-models/source.txt

@@ -0,0 +1,7 @@
+https://github.com/axinc-ai/ailia-models/tree/master/audio_processing/voicefilter
+
+https://storage.googleapis.com/ailia-models/voicefilter/embedder.onnx
+https://storage.googleapis.com/ailia-models/voicefilter/embedder.onnx.prototxt
+
+https://storage.googleapis.com/ailia-models/voicefilter/model.onnx
+https://storage.googleapis.com/ailia-models/voicefilter/model.onnx.prototxt

models/voicefilter/embedder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fb6732db7606dab53a37270d04057b25b54304aaf4f89281702bc5d7ba18d5f
+size 48540439

models/voicefilter/source.txt

@@ -0,0 +1,2 @@
+https://github.com/maum-ai/voicefilter
+https://drive.google.com/file/d/1YFmhmUok-W76JkrfA0fzQt3c-ZsfiwfL/view