Upload folder using huggingface_hub

.gitattributes

@@ -71,3 +71,4 @@ checkpoints-v4.6++/checkpoint-16384/eval_state.json filter=lfs diff=lfs merge=lf
 checkpoints-v4.6++/checkpoint-21504/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v5.14-b/checkpoint-11264/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v5.15/checkpoint-11264/eval_state.json filter=lfs diff=lfs merge=lfs -text
+checkpoints-v4.7/checkpoint-12288/eval_state.json filter=lfs diff=lfs merge=lfs -text

checkpoints-v4.7/checkpoint-12288/ema.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b4081e5e24b2250b7846093d661b30719021d18566a9784b71bf17bfb6a131d
+size 38935592

checkpoints-v4.7/checkpoint-12288/eval_state.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dee0e9731f3af014f9a423934ec72ede6d52f0404130329194aaed3b55d096f5
+size 43832293

checkpoints-v4.7/checkpoint-12288/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ffcc83093a481bd78981293c8db2808f3a75fe95581c3ac982ca4f69cc628701
+size 38935624

checkpoints-v4.7/checkpoint-12288/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15218a73c7b8b1f3e2f974b5e27a927346ea7a85770d930617ef2510c6c9caf6
+size 77992907

checkpoints-v4.7/checkpoint-12288/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:63f78f3ff75b9c12f2a6591a287446d50d5227658108c055e2b1960bf8289b9d
+size 14645

checkpoints-v4.7/checkpoint-12288/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b45d7713535c7997aee833eca7e2fe28a3aaf92d7b596cf836e30bb9f794858
+size 1383

checkpoints-v4.7/checkpoint-12288/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:362977a3fe3e0d6b4d1f69118ac28d15e3748dcb744d952cd08de6c9c27544b6
+size 1465

checkpoints-v4.7/checkpoint-12288/trainer_state.json

@@ -0,0 +1,790 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.5675488430095608,
+  "eval_steps": 1024,
+  "global_step": 12288,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011823934229365849,
+      "grad_norm": 0.8926580548286438,
+      "learning_rate": 0.000498046875,
+      "loss": 7.760124206542969,
+      "step": 256
+    },
+    {
+      "epoch": 0.023647868458731697,
+      "grad_norm": 1.5264979600906372,
+      "learning_rate": 0.000998046875,
+      "loss": 1.7502448558807373,
+      "step": 512
+    },
+    {
+      "epoch": 0.03547180268809755,
+      "grad_norm": 1.034440517425537,
+      "learning_rate": 0.000999640996023194,
+      "loss": 1.0943012237548828,
+      "step": 768
+    },
+    {
+      "epoch": 0.047295736917463395,
+      "grad_norm": 1.2952566146850586,
+      "learning_rate": 0.0009985588674043958,
+      "loss": 0.9299185276031494,
+      "step": 1024
+    },
+    {
+      "epoch": 0.047295736917463395,
+      "eval_bleu": 0.8108942681968176,
+      "eval_cos_loss": 0.18627598218313635,
+      "eval_dec_loss": 0.22058332814371476,
+      "eval_loss": 0.8581492928065122,
+      "eval_mse2_loss": 0.0729683047781388,
+      "eval_mse_loss": 0.5191411869302732,
+      "eval_rec_loss": 0.024248257404848185,
+      "eval_var_loss": 0.0025806165721318493,
+      "flow/cos_sim": 0.8137240362221791,
+      "flow/improvement_ratio": 0.953395878479361,
+      "flow/mag_ratio_mean": 0.8106630417309939,
+      "flow/mag_ratio_std": 0.14785969192714996,
+      "step": 1024
+    },
+    {
+      "epoch": 0.047295736917463395,
+      "eval_bleu": 0.8108942681968176,
+      "eval_cos_loss": 0.18627598218313635,
+      "eval_dec_loss": 0.22058332814371476,
+      "eval_loss": 0.8581492928065122,
+      "eval_mse2_loss": 0.0729683047781388,
+      "eval_mse_loss": 0.5191411869302732,
+      "eval_rec_loss": 0.024248257404848185,
+      "eval_runtime": 153.8108,
+      "eval_samples_per_second": 181.996,
+      "eval_steps_per_second": 2.848,
+      "eval_var_loss": 0.0025806165721318493,
+      "flow/cos_sim": 0.8137240362221791,
+      "flow/improvement_ratio": 0.953395878479361,
+      "flow/mag_ratio_mean": 0.8106630417309939,
+      "flow/mag_ratio_std": 0.14785969192714996,
+      "step": 1024
+    },
+    {
+      "epoch": 0.05911967114682925,
+      "grad_norm": 1.6030019521713257,
+      "learning_rate": 0.0009967551747861387,
+      "loss": 0.8445956110954285,
+      "step": 1280
+    },
+    {
+      "epoch": 0.0709436053761951,
+      "grad_norm": 1.6468690633773804,
+      "learning_rate": 0.000994232528651847,
+      "loss": 0.7937943339347839,
+      "step": 1536
+    },
+    {
+      "epoch": 0.08276753960556095,
+      "grad_norm": 1.9545633792877197,
+      "learning_rate": 0.0009909945800260092,
+      "loss": 0.7469877600669861,
+      "step": 1792
+    },
+    {
+      "epoch": 0.09459147383492679,
+      "grad_norm": 1.2975454330444336,
+      "learning_rate": 0.0009870460151900522,
+      "loss": 0.7277378439903259,
+      "step": 2048
+    },
+    {
+      "epoch": 0.09459147383492679,
+      "eval_bleu": 0.8332413753983823,
+      "eval_cos_loss": 0.12301718736212003,
+      "eval_dec_loss": 0.18318539169791354,
+      "eval_loss": 0.6979760440517234,
+      "eval_mse2_loss": 0.056748984509134945,
+      "eval_mse_loss": 0.43819310152095203,
+      "eval_rec_loss": 0.006146695995894708,
+      "eval_var_loss": 0.0014001506648651542,
+      "flow/cos_sim": 0.8769828294782334,
+      "flow/improvement_ratio": 0.9491692741984101,
+      "flow/mag_ratio_mean": 0.8741380685268472,
+      "flow/mag_ratio_std": 0.1099956316456675,
+      "step": 2048
+    },
+    {
+      "epoch": 0.09459147383492679,
+      "eval_bleu": 0.8332413753983823,
+      "eval_cos_loss": 0.12301718736212003,
+      "eval_dec_loss": 0.18318539169791354,
+      "eval_loss": 0.6979760440517234,
+      "eval_mse2_loss": 0.056748984509134945,
+      "eval_mse_loss": 0.43819310152095203,
+      "eval_rec_loss": 0.006146695995894708,
+      "eval_runtime": 147.7361,
+      "eval_samples_per_second": 189.48,
+      "eval_steps_per_second": 2.965,
+      "eval_var_loss": 0.0014001506648651542,
+      "flow/cos_sim": 0.8769828294782334,
+      "flow/improvement_ratio": 0.9491692741984101,
+      "flow/mag_ratio_mean": 0.8741380685268472,
+      "flow/mag_ratio_std": 0.1099956316456675,
+      "step": 2048
+    },
+    {
+      "epoch": 0.10641540806429264,
+      "grad_norm": 1.0802249908447266,
+      "learning_rate": 0.0009823925488998885,
+      "loss": 0.7090811133384705,
+      "step": 2304
+    },
+    {
+      "epoch": 0.1182393422936585,
+      "grad_norm": 1.3514765501022339,
+      "learning_rate": 0.0009770409161149525,
+      "loss": 0.6976003646850586,
+      "step": 2560
+    },
+    {
+      "epoch": 0.13006327652302435,
+      "grad_norm": 1.3863739967346191,
+      "learning_rate": 0.0009709988622506973,
+      "loss": 0.6783488392829895,
+      "step": 2816
+    },
+    {
+      "epoch": 0.1418872107523902,
+      "grad_norm": 1.2354881763458252,
+      "learning_rate": 0.000964275131968659,
+      "loss": 0.6734734773635864,
+      "step": 3072
+    },
+    {
+      "epoch": 0.1418872107523902,
+      "eval_bleu": 0.8490467382834135,
+      "eval_cos_loss": 0.09506861453867393,
+      "eval_dec_loss": 0.16530320855903707,
+      "eval_loss": 0.6531072377342068,
+      "eval_mse2_loss": 0.052244682587921344,
+      "eval_mse_loss": 0.4211963676426509,
+      "eval_rec_loss": 0.003093622522146896,
+      "eval_var_loss": 0.0017624946489726027,
+      "flow/cos_sim": 0.9049313971985421,
+      "flow/improvement_ratio": 0.950364352905587,
+      "flow/mag_ratio_mean": 0.8911371738671168,
+      "flow/mag_ratio_std": 0.10241742183764775,
+      "step": 3072
+    },
+    {
+      "epoch": 0.1418872107523902,
+      "eval_bleu": 0.8490467382834135,
+      "eval_cos_loss": 0.09506861453867393,
+      "eval_dec_loss": 0.16530320855903707,
+      "eval_loss": 0.6531072377342068,
+      "eval_mse2_loss": 0.052244682587921344,
+      "eval_mse_loss": 0.4211963676426509,
+      "eval_rec_loss": 0.003093622522146896,
+      "eval_runtime": 144.2539,
+      "eval_samples_per_second": 194.054,
+      "eval_steps_per_second": 3.036,
+      "eval_var_loss": 0.0017624946489726027,
+      "flow/cos_sim": 0.9049313971985421,
+      "flow/improvement_ratio": 0.950364352905587,
+      "flow/mag_ratio_mean": 0.8911371738671168,
+      "flow/mag_ratio_std": 0.10241742183764775,
+      "step": 3072
+    },
+    {
+      "epoch": 0.15371114498175603,
+      "grad_norm": 1.2442104816436768,
+      "learning_rate": 0.0009568794565203123,
+      "loss": 0.6588751673698425,
+      "step": 3328
+    },
+    {
+      "epoch": 0.1655350792111219,
+      "grad_norm": 1.032291054725647,
+      "learning_rate": 0.0009488225396630347,
+      "loss": 0.647803008556366,
+      "step": 3584
+    },
+    {
+      "epoch": 0.17735901344048774,
+      "grad_norm": 1.2258528470993042,
+      "learning_rate": 0.0009401160421685646,
+      "loss": 0.6433600783348083,
+      "step": 3840
+    },
+    {
+      "epoch": 0.18918294766985358,
+      "grad_norm": 1.1139024496078491,
+      "learning_rate": 0.0009307725649463714,
+      "loss": 0.6353314518928528,
+      "step": 4096
+    },
+    {
+      "epoch": 0.18918294766985358,
+      "eval_bleu": 0.8504080018321996,
+      "eval_cos_loss": 0.07373750197901029,
+      "eval_dec_loss": 0.15640341749066086,
+      "eval_loss": 0.6174904828610486,
+      "eval_mse2_loss": 0.04752455838024616,
+      "eval_mse_loss": 0.4026154523979039,
+      "eval_rec_loss": 0.002115985062733488,
+      "eval_var_loss": 0.0014573188677226027,
+      "flow/cos_sim": 0.9262625174979641,
+      "flow/improvement_ratio": 0.9461281280539352,
+      "flow/mag_ratio_mean": 0.9282138170716969,
+      "flow/mag_ratio_std": 0.0990274522844787,
+      "step": 4096
+    },
+    {
+      "epoch": 0.18918294766985358,
+      "eval_bleu": 0.8504080018321996,
+      "eval_cos_loss": 0.07373750197901029,
+      "eval_dec_loss": 0.15640341749066086,
+      "eval_loss": 0.6174904828610486,
+      "eval_mse2_loss": 0.04752455838024616,
+      "eval_mse_loss": 0.4026154523979039,
+      "eval_rec_loss": 0.002115985062733488,
+      "eval_runtime": 144.4539,
+      "eval_samples_per_second": 193.785,
+      "eval_steps_per_second": 3.032,
+      "eval_var_loss": 0.0014573188677226027,
+      "flow/cos_sim": 0.9262625174979641,
+      "flow/improvement_ratio": 0.9461281280539352,
+      "flow/mag_ratio_mean": 0.9282138170716969,
+      "flow/mag_ratio_std": 0.0990274522844787,
+      "step": 4096
+    },
+    {
+      "epoch": 0.20100688189921945,
+      "grad_norm": 1.0958069562911987,
+      "learning_rate": 0.0009208056308063659,
+      "loss": 0.6263965368270874,
+      "step": 4352
+    },
+    {
+      "epoch": 0.2128308161285853,
+      "grad_norm": 1.2000305652618408,
+      "learning_rate": 0.0009102296648873445,
+      "loss": 0.618091344833374,
+      "step": 4608
+    },
+    {
+      "epoch": 0.22465475035795113,
+      "grad_norm": 0.8440291285514832,
+      "learning_rate": 0.0008990599737794927,
+      "loss": 0.6174848079681396,
+      "step": 4864
+    },
+    {
+      "epoch": 0.236478684587317,
+      "grad_norm": 0.9181346893310547,
+      "learning_rate": 0.0008873127233711644,
+      "loss": 0.606370747089386,
+      "step": 5120
+    },
+    {
+      "epoch": 0.236478684587317,
+      "eval_bleu": 0.854693698336936,
+      "eval_cos_loss": 0.058853609810613064,
+      "eval_dec_loss": 0.15538246511188272,
+      "eval_loss": 0.5900675073334071,
+      "eval_mse2_loss": 0.04474926833488625,
+      "eval_mse_loss": 0.3815204039828418,
+      "eval_rec_loss": 0.0014702541673647402,
+      "eval_var_loss": 0.0010597542540667808,
+      "flow/cos_sim": 0.9411464126687071,
+      "flow/improvement_ratio": 0.9471585357298046,
+      "flow/mag_ratio_mean": 0.9485053326169105,
+      "flow/mag_ratio_std": 0.0952613887347315,
+      "step": 5120
+    },
+    {
+      "epoch": 0.236478684587317,
+      "eval_bleu": 0.854693698336936,
+      "eval_cos_loss": 0.058853609810613064,
+      "eval_dec_loss": 0.15538246511188272,
+      "eval_loss": 0.5900675073334071,
+      "eval_mse2_loss": 0.04474926833488625,
+      "eval_mse_loss": 0.3815204039828418,
+      "eval_rec_loss": 0.0014702541673647402,
+      "eval_runtime": 144.8825,
+      "eval_samples_per_second": 193.212,
+      "eval_steps_per_second": 3.023,
+      "eval_var_loss": 0.0010597542540667808,
+      "flow/cos_sim": 0.9411464126687071,
+      "flow/improvement_ratio": 0.9471585357298046,
+      "flow/mag_ratio_mean": 0.9485053326169105,
+      "flow/mag_ratio_std": 0.0952613887347315,
+      "step": 5120
+    },
+    {
+      "epoch": 0.24830261881668284,
+      "grad_norm": 1.0105012655258179,
+      "learning_rate": 0.0008750049154520011,
+      "loss": 0.5982246398925781,
+      "step": 5376
+    },
+    {
+      "epoch": 0.2601265530460487,
+      "grad_norm": 1.7041122913360596,
+      "learning_rate": 0.0008621543631062487,
+      "loss": 0.6011320948600769,
+      "step": 5632
+    },
+    {
+      "epoch": 0.27195048727541454,
+      "grad_norm": 1.2530128955841064,
+      "learning_rate": 0.0008487796649318904,
+      "loss": 0.5908714532852173,
+      "step": 5888
+    },
+    {
+      "epoch": 0.2837744215047804,
+      "grad_norm": 0.8214261531829834,
+      "learning_rate": 0.0008349001781229053,
+      "loss": 0.5860975980758667,
+      "step": 6144
+    },
+    {
+      "epoch": 0.2837744215047804,
+      "eval_bleu": 0.8588448853802904,
+      "eval_cos_loss": 0.050389231161371756,
+      "eval_dec_loss": 0.14640251302147564,
+      "eval_loss": 0.5701193273339642,
+      "eval_mse2_loss": 0.04257897145569869,
+      "eval_mse_loss": 0.3735292319837771,
+      "eval_rec_loss": 0.0011887007649628058,
+      "eval_var_loss": 0.001380990084992152,
+      "flow/cos_sim": 0.9496107875756478,
+      "flow/improvement_ratio": 0.9471838724667623,
+      "flow/mag_ratio_mean": 0.9603897756365336,
+      "flow/mag_ratio_std": 0.09184158112020253,
+      "step": 6144
+    },
+    {
+      "epoch": 0.2837744215047804,
+      "eval_bleu": 0.8588448853802904,
+      "eval_cos_loss": 0.050389231161371756,
+      "eval_dec_loss": 0.14640251302147564,
+      "eval_loss": 0.5701193273339642,
+      "eval_mse2_loss": 0.04257897145569869,
+      "eval_mse_loss": 0.3735292319837771,
+      "eval_rec_loss": 0.0011887007649628058,
+      "eval_runtime": 145.6643,
+      "eval_samples_per_second": 192.175,
+      "eval_steps_per_second": 3.007,
+      "eval_var_loss": 0.001380990084992152,
+      "flow/cos_sim": 0.9496107875756478,
+      "flow/improvement_ratio": 0.9471838724667623,
+      "flow/mag_ratio_mean": 0.9603897756365336,
+      "flow/mag_ratio_std": 0.09184158112020253,
+      "step": 6144
+    },
+    {
+      "epoch": 0.2955983557341462,
+      "grad_norm": 0.9015347361564636,
+      "learning_rate": 0.0008205359904536107,
+      "loss": 0.5818743705749512,
+      "step": 6400
+    },
+    {
+      "epoch": 0.30742228996351206,
+      "grad_norm": 0.8061195611953735,
+      "learning_rate": 0.0008057078912056363,
+      "loss": 0.5712096691131592,
+      "step": 6656
+    },
+    {
+      "epoch": 0.3192462241928779,
+      "grad_norm": 0.9496876001358032,
+      "learning_rate": 0.0007904373410796086,
+      "loss": 0.5827493667602539,
+      "step": 6912
+    },
+    {
+      "epoch": 0.3310701584222438,
+      "grad_norm": 0.9318349957466125,
+      "learning_rate": 0.0007747464411350876,
+      "loss": 0.5713083744049072,
+      "step": 7168
+    },
+    {
+      "epoch": 0.3310701584222438,
+      "eval_bleu": 0.8655904613504272,
+      "eval_cos_loss": 0.04345733123793177,
+      "eval_dec_loss": 0.1383160431550382,
+      "eval_loss": 0.560146918988119,
+      "eval_mse2_loss": 0.04110090016093973,
+      "eval_mse_loss": 0.37529142364247203,
+      "eval_rec_loss": 0.0010027640903040077,
+      "eval_var_loss": 9.005472540311074e-05,
+      "flow/cos_sim": 0.956542692227995,
+      "flow/improvement_ratio": 0.9472023688222719,
+      "flow/mag_ratio_mean": 0.9649291216782784,
+      "flow/mag_ratio_std": 0.08745992025488043,
+      "step": 7168
+    },
+    {
+      "epoch": 0.3310701584222438,
+      "eval_bleu": 0.8655904613504272,
+      "eval_cos_loss": 0.04345733123793177,
+      "eval_dec_loss": 0.1383160431550382,
+      "eval_loss": 0.560146918988119,
+      "eval_mse2_loss": 0.04110090016093973,
+      "eval_mse_loss": 0.37529142364247203,
+      "eval_rec_loss": 0.0010027640903040077,
+      "eval_runtime": 147.3557,
+      "eval_samples_per_second": 189.969,
+      "eval_steps_per_second": 2.972,
+      "eval_var_loss": 9.005472540311074e-05,
+      "flow/cos_sim": 0.956542692227995,
+      "flow/improvement_ratio": 0.9472023688222719,
+      "flow/mag_ratio_mean": 0.9649291216782784,
+      "flow/mag_ratio_std": 0.08745992025488043,
+      "step": 7168
+    },
+    {
+      "epoch": 0.34289409265160964,
+      "grad_norm": 1.0284937620162964,
+      "learning_rate": 0.000758657900803716,
+      "loss": 0.5740544199943542,
+      "step": 7424
+    },
+    {
+      "epoch": 0.3547180268809755,
+      "grad_norm": 0.7206848859786987,
+      "learning_rate": 0.000742195005021869,
+      "loss": 0.5760706067085266,
+      "step": 7680
+    },
+    {
+      "epoch": 0.3665419611103413,
+      "grad_norm": 1.221917748451233,
+      "learning_rate": 0.0007253815805303786,
+      "loss": 0.566294252872467,
+      "step": 7936
+    },
+    {
+      "epoch": 0.37836589533970716,
+      "grad_norm": 0.8819605708122253,
+      "learning_rate": 0.0007082419613901028,
+      "loss": 0.5628067851066589,
+      "step": 8192
+    },
+    {
+      "epoch": 0.37836589533970716,
+      "eval_bleu": 0.871946148436791,
+      "eval_cos_loss": 0.03890436059331785,
+      "eval_dec_loss": 0.1332924633104031,
+      "eval_loss": 0.5495835452183196,
+      "eval_mse2_loss": 0.040371610187674496,
+      "eval_mse_loss": 0.3710057751773155,
+      "eval_rec_loss": 0.0008419993847885518,
+      "eval_var_loss": 0.0001812590855986016,
+      "flow/cos_sim": 0.9610956558898159,
+      "flow/improvement_ratio": 0.9478674083267717,
+      "flow/mag_ratio_mean": 0.9691984739205609,
+      "flow/mag_ratio_std": 0.08465842693431737,
+      "step": 8192
+    },
+    {
+      "epoch": 0.37836589533970716,
+      "eval_bleu": 0.871946148436791,
+      "eval_cos_loss": 0.03890436059331785,
+      "eval_dec_loss": 0.1332924633104031,
+      "eval_loss": 0.5495835452183196,
+      "eval_mse2_loss": 0.040371610187674496,
+      "eval_mse_loss": 0.3710057751773155,
+      "eval_rec_loss": 0.0008419993847885518,
+      "eval_runtime": 146.7514,
+      "eval_samples_per_second": 190.751,
+      "eval_steps_per_second": 2.985,
+      "eval_var_loss": 0.0001812590855986016,
+      "flow/cos_sim": 0.9610956558898159,
+      "flow/improvement_ratio": 0.9478674083267717,
+      "flow/mag_ratio_mean": 0.9691984739205609,
+      "flow/mag_ratio_std": 0.08465842693431737,
+      "step": 8192
+    },
+    {
+      "epoch": 0.390189829569073,
+      "grad_norm": 1.2066078186035156,
+      "learning_rate": 0.0006908009537632514,
+      "loss": 0.5644704699516296,
+      "step": 8448
+    },
+    {
+      "epoch": 0.4020137637984389,
+      "grad_norm": 1.2743791341781616,
+      "learning_rate": 0.0006730838000114403,
+      "loss": 0.5624759197235107,
+      "step": 8704
+    },
+    {
+      "epoch": 0.41383769802780473,
+      "grad_norm": 0.6424040198326111,
+      "learning_rate": 0.0006551161421624341,
+      "loss": 0.5654159188270569,
+      "step": 8960
+    },
+    {
+      "epoch": 0.4256616322571706,
+      "grad_norm": 1.0390995740890503,
+      "learning_rate": 0.0006369239847984517,
+      "loss": 0.5563592910766602,
+      "step": 9216
+    },
+    {
+      "epoch": 0.4256616322571706,
+      "eval_bleu": 0.8641208937990813,
+      "eval_cos_loss": 0.034672807719235275,
+      "eval_dec_loss": 0.14222022919962396,
+      "eval_loss": 0.5460346293503835,
+      "eval_mse2_loss": 0.03956130868223704,
+      "eval_mse_loss": 0.3584019422667212,
+      "eval_rec_loss": 0.0007127871282784625,
+      "eval_var_loss": 0.0016710812642694063,
+      "flow/cos_sim": 0.965327212663546,
+      "flow/improvement_ratio": 0.946786184561307,
+      "flow/mag_ratio_mean": 0.9734560983366074,
+      "flow/mag_ratio_std": 0.08033725547872178,
+      "step": 9216
+    },
+    {
+      "epoch": 0.4256616322571706,
+      "eval_bleu": 0.8641208937990813,
+      "eval_cos_loss": 0.034672807719235275,
+      "eval_dec_loss": 0.14222022919962396,
+      "eval_loss": 0.5460346293503835,
+      "eval_mse2_loss": 0.03956130868223704,
+      "eval_mse_loss": 0.3584019422667212,
+      "eval_rec_loss": 0.0007127871282784625,
+      "eval_runtime": 146.177,
+      "eval_samples_per_second": 191.501,
+      "eval_steps_per_second": 2.996,
+      "eval_var_loss": 0.0016710812642694063,
+      "flow/cos_sim": 0.965327212663546,
+      "flow/improvement_ratio": 0.946786184561307,
+      "flow/mag_ratio_mean": 0.9734560983366074,
+      "flow/mag_ratio_std": 0.08033725547872178,
+      "step": 9216
+    },
+    {
+      "epoch": 0.4374855664865364,
+      "grad_norm": 1.0673410892486572,
+      "learning_rate": 0.0006185336574197479,
+      "loss": 0.55131596326828,
+      "step": 9472
+    },
+    {
+      "epoch": 0.44930950071590225,
+      "grad_norm": 1.1385674476623535,
+      "learning_rate": 0.0005999717763379407,
+      "loss": 0.5542811155319214,
+      "step": 9728
+    },
+    {
+      "epoch": 0.4611334349452681,
+      "grad_norm": 1.3084577322006226,
+      "learning_rate": 0.0005812652061542363,
+      "loss": 0.5522482395172119,
+      "step": 9984
+    },
+    {
+      "epoch": 0.472957369174634,
+      "grad_norm": 0.9078991413116455,
+      "learning_rate": 0.0005624410208783071,
+      "loss": 0.5514112114906311,
+      "step": 10240
+    },
+    {
+      "epoch": 0.472957369174634,
+      "eval_bleu": 0.8714027910618674,
+      "eval_cos_loss": 0.031159216023505278,
+      "eval_dec_loss": 0.13065314148744084,
+      "eval_loss": 0.5320979358126584,
+      "eval_mse2_loss": 0.038692950696331455,
+      "eval_mse_loss": 0.3575415439121255,
+      "eval_rec_loss": 0.0006259792361515585,
+      "eval_var_loss": 0.001468397166630993,
+      "flow/cos_sim": 0.968840807404148,
+      "flow/improvement_ratio": 0.9490241132098246,
+      "flow/mag_ratio_mean": 0.9646270977307673,
+      "flow/mag_ratio_std": 0.07568225521407171,
+      "step": 10240
+    },
+    {
+      "epoch": 0.472957369174634,
+      "eval_bleu": 0.8714027910618674,
+      "eval_cos_loss": 0.031159216023505278,
+      "eval_dec_loss": 0.13065314148744084,
+      "eval_loss": 0.5320979358126584,
+      "eval_mse2_loss": 0.038692950696331455,
+      "eval_mse_loss": 0.3575415439121255,
+      "eval_rec_loss": 0.0006259792361515585,
+      "eval_runtime": 147.3748,
+      "eval_samples_per_second": 189.944,
+      "eval_steps_per_second": 2.972,
+      "eval_var_loss": 0.001468397166630993,
+      "flow/cos_sim": 0.968840807404148,
+      "flow/improvement_ratio": 0.9490241132098246,
+      "flow/mag_ratio_mean": 0.9646270977307673,
+      "flow/mag_ratio_std": 0.07568225521407171,
+      "step": 10240
+    },
+    {
+      "epoch": 0.48478130340399983,
+      "grad_norm": 1.3058608770370483,
+      "learning_rate": 0.0005435264647440881,
+      "loss": 0.547296941280365,
+      "step": 10496
+    },
+    {
+      "epoch": 0.49660523763336567,
+      "grad_norm": 1.0200841426849365,
+      "learning_rate": 0.000524548912779213,
+      "loss": 0.544040322303772,
+      "step": 10752
+    },
+    {
+      "epoch": 0.5084291718627315,
+      "grad_norm": 1.1076935529708862,
+      "learning_rate": 0.0005055358311851499,
+      "loss": 0.5454155206680298,
+      "step": 11008
+    },
+    {
+      "epoch": 0.5202531060920974,
+      "grad_norm": 0.8338369727134705,
+      "learning_rate": 0.0004865147375853812,
+      "loss": 0.5434398651123047,
+      "step": 11264
+    },
+    {
+      "epoch": 0.5202531060920974,
+      "eval_bleu": 0.8728636919662941,
+      "eval_cos_loss": 0.029134724701682456,
+      "eval_dec_loss": 0.130046037353201,
+      "eval_loss": 0.5287471506829675,
+      "eval_mse2_loss": 0.037880827574969425,
+      "eval_mse_loss": 0.3566115467243543,
+      "eval_rec_loss": 0.0005072098316288007,
+      "eval_var_loss": 0.0007880572314676084,
+      "flow/cos_sim": 0.9708652933166452,
+      "flow/improvement_ratio": 0.9513337348149792,
+      "flow/mag_ratio_mean": 0.9743897437232815,
+      "flow/mag_ratio_std": 0.07394225260914733,
+      "step": 11264
+    },
+    {
+      "epoch": 0.5202531060920974,
+      "eval_bleu": 0.8728636919662941,
+      "eval_cos_loss": 0.029134724701682456,
+      "eval_dec_loss": 0.130046037353201,
+      "eval_loss": 0.5287471506829675,
+      "eval_mse2_loss": 0.037880827574969425,
+      "eval_mse_loss": 0.3566115467243543,
+      "eval_rec_loss": 0.0005072098316288007,
+      "eval_runtime": 147.0551,
+      "eval_samples_per_second": 190.357,
+      "eval_steps_per_second": 2.978,
+      "eval_var_loss": 0.0007880572314676084,
+      "flow/cos_sim": 0.9708652933166452,
+      "flow/improvement_ratio": 0.9513337348149792,
+      "flow/mag_ratio_mean": 0.9743897437232815,
+      "flow/mag_ratio_std": 0.07394225260914733,
+      "step": 11264
+    },
+    {
+      "epoch": 0.5320770403214632,
+      "grad_norm": 1.2937544584274292,
+      "learning_rate": 0.0004675131611991607,
+      "loss": 0.5423741936683655,
+      "step": 11520
+    },
+    {
+      "epoch": 0.5439009745508291,
+      "grad_norm": 0.9950433373451233,
+      "learning_rate": 0.0004485586029984899,
+      "loss": 0.536012589931488,
+      "step": 11776
+    },
+    {
+      "epoch": 0.5557249087801949,
+      "grad_norm": 0.9091536402702332,
+      "learning_rate": 0.00042967849590597266,
+      "loss": 0.5319453477859497,
+      "step": 12032
+    },
+    {
+      "epoch": 0.5675488430095608,
+      "grad_norm": 1.1773775815963745,
+      "learning_rate": 0.0004109001650911621,
+      "loss": 0.5343883037567139,
+      "step": 12288
+    },
+    {
+      "epoch": 0.5675488430095608,
+      "eval_bleu": 0.8703424178205617,
+      "eval_cos_loss": 0.026759936142559736,
+      "eval_dec_loss": 0.1364484317855884,
+      "eval_loss": 0.5287964605305293,
+      "eval_mse2_loss": 0.03801968024796955,
+      "eval_mse_loss": 0.34965787868793696,
+      "eval_rec_loss": 0.00048410188376538874,
+      "eval_var_loss": 0.0015103762552618437,
+      "flow/cos_sim": 0.9732400857966784,
+      "flow/improvement_ratio": 0.9476525126254722,
+      "flow/mag_ratio_mean": 0.9758574831975649,
+      "flow/mag_ratio_std": 0.06964536890659702,
+      "step": 12288
+    },
+    {
+      "epoch": 0.5675488430095608,
+      "eval_bleu": 0.8703424178205617,
+      "eval_cos_loss": 0.026759936142559736,
+      "eval_dec_loss": 0.1364484317855884,
+      "eval_loss": 0.5287964605305293,
+      "eval_mse2_loss": 0.03801968024796955,
+      "eval_mse_loss": 0.34965787868793696,
+      "eval_rec_loss": 0.00048410188376538874,
+      "eval_runtime": 146.8515,
+      "eval_samples_per_second": 190.621,
+      "eval_steps_per_second": 2.983,
+      "eval_var_loss": 0.0015103762552618437,
+      "flow/cos_sim": 0.9732400857966784,
+      "flow/improvement_ratio": 0.9476525126254722,
+      "flow/mag_ratio_mean": 0.9758574831975649,
+      "flow/mag_ratio_std": 0.06964536890659702,
+      "step": 12288
+    }
+  ],
+  "logging_steps": 256,
+  "max_steps": 21651,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 1024,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints-v4.7/checkpoint-12288/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61a16bb839f687414b8e48611327c4b9cfddeefe38c031ca70808f9a97c476b7
+size 5137