A

config.json

@@ -1,6 +1,6 @@
 {
-  "_name_or_path": "lighteternal/wav2vec2-large-xlsr-53-greek",
-  "activation_dropout": 0.0,
+  "_name_or_path": "jonatasgrosman/wav2vec2-large-xlsr-53-english",
+  "activation_dropout": 0.05,
   "adapter_kernel_size": 3,
   "adapter_stride": 2,
   "add_adapter": false,
@@ -49,12 +49,12 @@
   "feat_extract_activation": "gelu",
   "feat_extract_dropout": 0.0,
   "feat_extract_norm": "layer",
-  "feat_proj_dropout": 0.0,
+  "feat_proj_dropout": 0.05,
   "feat_quantizer_dropout": 0.0,
   "final_dropout": 0.0,
   "finetuning_task": "wav2vec2_clf",
   "hidden_act": "gelu",
-  "hidden_dropout": 0.1,
+  "hidden_dropout": 0.05,
   "hidden_size": 1024,
   "id2label": {
     "0": "angry",
@@ -79,7 +79,7 @@
     "surprised": 7
   },
   "layer_norm_eps": 1e-05,
-  "layerdrop": 0.1,
+  "layerdrop": 0.05,
   "mask_channel_length": 10,
   "mask_channel_min_space": 1,
   "mask_channel_other": 0.0,
@@ -105,7 +105,7 @@
   "num_hidden_layers": 24,
   "num_negatives": 100,
   "output_hidden_size": 1024,
-  "pad_token_id": 54,
+  "pad_token_id": 0,
   "pooling_mode": "mean",
   "problem_type": "single_label_classification",
   "proj_codevector_dim": 256,
@@ -133,6 +133,6 @@
   "torch_dtype": "float32",
   "transformers_version": "4.23.0.dev0",
   "use_weighted_layer_sum": false,
-  "vocab_size": 55,
+  "vocab_size": 33,
   "xvector_output_dim": 512
 }

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e2ebb6b6119110351be64f39f20826294fb7a16faeb39b79f1e56721d08f2ca7
-size 2498522121
+oid sha256:3175558e8e61233286d0233755e426ebc33a4a0ee9d88efeabc91b0969dd0d6c
+size 2498522505

preprocessor_config.json

@@ -4,6 +4,7 @@
   "feature_size": 1,
   "padding_side": "right",
   "padding_value": 0.0,
+  "processor_class": "Wav2Vec2ProcessorWithLM",
   "return_attention_mask": true,
   "sampling_rate": 16000
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:21fe72d43c31b3044b1ec946fa3b32078ed99dea846cd0192a26856bb9c839e6
+oid sha256:b83378c33a6ea5b49584394b6889d866dda8704d4223b7e012cbbcb8bc2469dc
 size 1266130541

rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f449ee5f1d7ffb1d2e2b47df15b3e8c0e25a55972ba52ec93a6b01f10b6ecec
-size 14567
+oid sha256:d2d6c598ad1ff03fdbae199c042313f72c379e59a7adc5c986920487ec0ef0f0
+size 14503

runs/Sep19_09-50-03_gpu-vision-2/1663581018.9298472/events.out.tfevents.1663581018.gpu-vision-2.7343.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3140c2710db4566bf2ad01dbb3a62bac7c2687c39b7d9f949ae7287b6185ded
+size 5619

runs/Sep19_09-50-03_gpu-vision-2/1663581051.79942/events.out.tfevents.1663581051.gpu-vision-2.7343.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4464bcac36593f7ef90c8eca0de8c4702da993d236d35ff3bd45c135b321f22d
+size 5619

runs/Sep19_09-50-03_gpu-vision-2/1663581091.68026/events.out.tfevents.1663581091.gpu-vision-2.7343.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0feff904f9fed500bff8dee84dc0234472bc8b4e6d6a28fa76bed2a005f7c9f3
+size 5619

runs/Sep19_09-50-03_gpu-vision-2/events.out.tfevents.1663581018.gpu-vision-2.7343.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8789fe19991f2241250a9dbd8d34b3ce8e51d6b2da673a01d4775562f69273b
+size 16424

runs/Sep19_09-54-01_gpu-vision-2/1663581250.7896519/events.out.tfevents.1663581250.gpu-vision-2.8089.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18097009ac52a860710c3cd723b1a48a261cc1608031261ab392c9f95de66a39
+size 5619

runs/Sep19_09-54-01_gpu-vision-2/events.out.tfevents.1663581250.gpu-vision-2.8089.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9c72015b89a11731dacf7c862046d9e79a7e5b2e1a2c81230f551cad4158f0c
+size 5811

runs/Sep19_09-57-55_gpu-vision-2/1663581500.0802476/events.out.tfevents.1663581500.gpu-vision-2.8798.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f877883a78a6de8ee3d243ce3aced488c2d6fcd75dd326dcc366584350517ab
+size 5619

runs/Sep19_09-57-55_gpu-vision-2/events.out.tfevents.1663581500.gpu-vision-2.8798.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13955865b2c38946709f3166fcee2f2435c98d539e38bf7986437a9d12002422
+size 5823

runs/Sep19_10-01-26_gpu-vision-2/1663581692.4217086/events.out.tfevents.1663581692.gpu-vision-2.8798.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e16446a28f142ef2d65c4e3450621714001bfc37f416d44d8370026f3826b99
+size 5619

runs/Sep19_10-01-26_gpu-vision-2/events.out.tfevents.1663581692.gpu-vision-2.8798.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa7f3fb3b39f40912afe701a7d7866f1bc9d46534613c425fb52101f5e21dfe9
+size 5823

runs/Sep19_10-07-32_gpu-vision-2/1663582066.1218033/events.out.tfevents.1663582066.gpu-vision-2.10212.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:252c9367a243110d4ed1a0e8f09eca4dc94235896b81d07ec241e7503a89b64c
+size 5619

runs/Sep19_10-07-32_gpu-vision-2/events.out.tfevents.1663582066.gpu-vision-2.10212.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbc41b733620722f3431a77db35789f4279e9c176417a1e65f4cd597022f743f
+size 41392

scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a5340902b17e35966e7e8d13c891b4a72edbb8cbd01bf429032179aa2bf58052
+oid sha256:a4b9accfbdcb85276435ad005655da973cd6905daf1d6df04732beae9fcc5608
 size 559

scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a9112768c193dcd7407fa487c3348c49ebdeabbef697849491bcfab0bd57910e
+oid sha256:0408533f367f94f43315c4942763b084245dd96c78cefc95eb3b423200fa39bb
 size 623

trainer_state.json

@@ -1,91 +1,1126 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.5,
-  "global_step": 50,
+  "epoch": 2.5694444444444446,
+  "global_step": 740,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.1,
-      "learning_rate": 0.0,
-      "loss": 1.8457,
+      "epoch": 0.03,
+      "learning_rate": 9.88425925925926e-05,
+      "loss": 2.0821,
       "step": 10
     },
     {
-      "epoch": 0.1,
-      "eval_accuracy": 0.23000000417232513,
-      "eval_loss": 1.845852017402649,
-      "eval_runtime": 42.6127,
-      "eval_samples_per_second": 4.693,
-      "eval_steps_per_second": 1.173,
+      "epoch": 0.03,
+      "eval_accuracy": 0.1527777761220932,
+      "eval_loss": 2.099255084991455,
+      "eval_runtime": 90.7163,
+      "eval_samples_per_second": 6.349,
+      "eval_steps_per_second": 1.587,
       "step": 10
     },
     {
-      "epoch": 0.2,
-      "learning_rate": 0.0,
-      "loss": 1.9441,
+      "epoch": 0.07,
+      "learning_rate": 9.768518518518519e-05,
+      "loss": 1.9991,
       "step": 20
     },
     {
-      "epoch": 0.2,
-      "eval_accuracy": 0.23000000417232513,
-      "eval_loss": 1.845852017402649,
-      "eval_runtime": 42.6438,
-      "eval_samples_per_second": 4.69,
-      "eval_steps_per_second": 1.173,
+      "epoch": 0.07,
+      "eval_accuracy": 0.2326388955116272,
+      "eval_loss": 2.011439085006714,
+      "eval_runtime": 90.9685,
+      "eval_samples_per_second": 6.332,
+      "eval_steps_per_second": 1.583,
       "step": 20
     },
     {
-      "epoch": 0.3,
-      "learning_rate": 0.0,
-      "loss": 1.8242,
+      "epoch": 0.1,
+      "learning_rate": 9.664351851851853e-05,
+      "loss": 2.0133,
       "step": 30
     },
     {
-      "epoch": 0.3,
-      "eval_accuracy": 0.23000000417232513,
-      "eval_loss": 1.845852017402649,
-      "eval_runtime": 42.8454,
-      "eval_samples_per_second": 4.668,
-      "eval_steps_per_second": 1.167,
+      "epoch": 0.1,
+      "eval_accuracy": 0.1822916716337204,
+      "eval_loss": 2.0611398220062256,
+      "eval_runtime": 91.1422,
+      "eval_samples_per_second": 6.32,
+      "eval_steps_per_second": 1.58,
       "step": 30
     },
     {
-      "epoch": 0.4,
-      "learning_rate": 0.0,
-      "loss": 1.8451,
+      "epoch": 0.14,
+      "learning_rate": 9.548611111111112e-05,
+      "loss": 1.9912,
       "step": 40
     },
     {
-      "epoch": 0.4,
-      "eval_accuracy": 0.23000000417232513,
-      "eval_loss": 1.845852017402649,
-      "eval_runtime": 42.5518,
-      "eval_samples_per_second": 4.7,
-      "eval_steps_per_second": 1.175,
+      "epoch": 0.14,
+      "eval_accuracy": 0.1944444477558136,
+      "eval_loss": 1.9874335527420044,
+      "eval_runtime": 91.2497,
+      "eval_samples_per_second": 6.312,
+      "eval_steps_per_second": 1.578,
       "step": 40
     },
     {
-      "epoch": 0.5,
-      "learning_rate": 0.0,
-      "loss": 1.7465,
+      "epoch": 0.17,
+      "learning_rate": 9.432870370370372e-05,
+      "loss": 1.9825,
       "step": 50
     },
     {
-      "epoch": 0.5,
-      "eval_accuracy": 0.23000000417232513,
-      "eval_loss": 1.845852017402649,
-      "eval_runtime": 42.3785,
-      "eval_samples_per_second": 4.719,
-      "eval_steps_per_second": 1.18,
+      "epoch": 0.17,
+      "eval_accuracy": 0.1875,
+      "eval_loss": 1.9108080863952637,
+      "eval_runtime": 91.363,
+      "eval_samples_per_second": 6.305,
+      "eval_steps_per_second": 1.576,
       "step": 50
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.31712962962963e-05,
+      "loss": 1.8281,
+      "step": 60
+    },
+    {
+      "epoch": 0.21,
+      "eval_accuracy": 0.2760416567325592,
+      "eval_loss": 1.8094313144683838,
+      "eval_runtime": 90.7999,
+      "eval_samples_per_second": 6.344,
+      "eval_steps_per_second": 1.586,
+      "step": 60
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.201388888888889e-05,
+      "loss": 1.7768,
+      "step": 70
+    },
+    {
+      "epoch": 0.24,
+      "eval_accuracy": 0.28125,
+      "eval_loss": 1.8212822675704956,
+      "eval_runtime": 90.3717,
+      "eval_samples_per_second": 6.374,
+      "eval_steps_per_second": 1.593,
+      "step": 70
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.085648148148149e-05,
+      "loss": 1.8747,
+      "step": 80
+    },
+    {
+      "epoch": 0.28,
+      "eval_accuracy": 0.2916666567325592,
+      "eval_loss": 1.819143533706665,
+      "eval_runtime": 89.6205,
+      "eval_samples_per_second": 6.427,
+      "eval_steps_per_second": 1.607,
+      "step": 80
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.969907407407407e-05,
+      "loss": 1.8258,
+      "step": 90
+    },
+    {
+      "epoch": 0.31,
+      "eval_accuracy": 0.2673611044883728,
+      "eval_loss": 1.8833109140396118,
+      "eval_runtime": 90.3547,
+      "eval_samples_per_second": 6.375,
+      "eval_steps_per_second": 1.594,
+      "step": 90
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.854166666666667e-05,
+      "loss": 1.8466,
+      "step": 100
+    },
+    {
+      "epoch": 0.35,
+      "eval_accuracy": 0.3020833432674408,
+      "eval_loss": 1.794202208518982,
+      "eval_runtime": 89.9211,
+      "eval_samples_per_second": 6.406,
+      "eval_steps_per_second": 1.601,
+      "step": 100
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.738425925925926e-05,
+      "loss": 1.7949,
+      "step": 110
+    },
+    {
+      "epoch": 0.38,
+      "eval_accuracy": 0.234375,
+      "eval_loss": 1.737123727798462,
+      "eval_runtime": 90.2733,
+      "eval_samples_per_second": 6.381,
+      "eval_steps_per_second": 1.595,
+      "step": 110
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 8.622685185185186e-05,
+      "loss": 1.6993,
+      "step": 120
+    },
+    {
+      "epoch": 0.42,
+      "eval_accuracy": 0.3333333432674408,
+      "eval_loss": 1.714468240737915,
+      "eval_runtime": 90.6521,
+      "eval_samples_per_second": 6.354,
+      "eval_steps_per_second": 1.588,
+      "step": 120
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 8.506944444444444e-05,
+      "loss": 1.9949,
+      "step": 130
+    },
+    {
+      "epoch": 0.45,
+      "eval_accuracy": 0.3628472089767456,
+      "eval_loss": 1.7169595956802368,
+      "eval_runtime": 89.657,
+      "eval_samples_per_second": 6.424,
+      "eval_steps_per_second": 1.606,
+      "step": 130
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 8.391203703703704e-05,
+      "loss": 1.6402,
+      "step": 140
+    },
+    {
+      "epoch": 0.49,
+      "eval_accuracy": 0.3506944477558136,
+      "eval_loss": 1.7289636135101318,
+      "eval_runtime": 90.3787,
+      "eval_samples_per_second": 6.373,
+      "eval_steps_per_second": 1.593,
+      "step": 140
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 8.275462962962963e-05,
+      "loss": 1.7599,
+      "step": 150
+    },
+    {
+      "epoch": 0.52,
+      "eval_accuracy": 0.3090277910232544,
+      "eval_loss": 1.7477116584777832,
+      "eval_runtime": 90.5736,
+      "eval_samples_per_second": 6.359,
+      "eval_steps_per_second": 1.59,
+      "step": 150
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.159722222222223e-05,
+      "loss": 1.5776,
+      "step": 160
+    },
+    {
+      "epoch": 0.56,
+      "eval_accuracy": 0.3715277910232544,
+      "eval_loss": 1.6158109903335571,
+      "eval_runtime": 89.7843,
+      "eval_samples_per_second": 6.415,
+      "eval_steps_per_second": 1.604,
+      "step": 160
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.043981481481482e-05,
+      "loss": 1.7169,
+      "step": 170
+    },
+    {
+      "epoch": 0.59,
+      "eval_accuracy": 0.3663194477558136,
+      "eval_loss": 1.6075005531311035,
+      "eval_runtime": 90.228,
+      "eval_samples_per_second": 6.384,
+      "eval_steps_per_second": 1.596,
+      "step": 170
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 7.928240740740742e-05,
+      "loss": 1.653,
+      "step": 180
+    },
+    {
+      "epoch": 0.62,
+      "eval_accuracy": 0.4201388955116272,
+      "eval_loss": 1.524334192276001,
+      "eval_runtime": 91.575,
+      "eval_samples_per_second": 6.29,
+      "eval_steps_per_second": 1.572,
+      "step": 180
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 7.8125e-05,
+      "loss": 1.5733,
+      "step": 190
+    },
+    {
+      "epoch": 0.66,
+      "eval_accuracy": 0.359375,
+      "eval_loss": 1.7072927951812744,
+      "eval_runtime": 90.2896,
+      "eval_samples_per_second": 6.379,
+      "eval_steps_per_second": 1.595,
+      "step": 190
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 7.69675925925926e-05,
+      "loss": 1.6704,
+      "step": 200
+    },
+    {
+      "epoch": 0.69,
+      "eval_accuracy": 0.4722222089767456,
+      "eval_loss": 1.4267817735671997,
+      "eval_runtime": 90.57,
+      "eval_samples_per_second": 6.36,
+      "eval_steps_per_second": 1.59,
+      "step": 200
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 7.581018518518519e-05,
+      "loss": 1.4389,
+      "step": 210
+    },
+    {
+      "epoch": 0.73,
+      "eval_accuracy": 0.3940972089767456,
+      "eval_loss": 1.7218824625015259,
+      "eval_runtime": 89.8269,
+      "eval_samples_per_second": 6.412,
+      "eval_steps_per_second": 1.603,
+      "step": 210
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 7.465277777777779e-05,
+      "loss": 1.5342,
+      "step": 220
+    },
+    {
+      "epoch": 0.76,
+      "eval_accuracy": 0.3940972089767456,
+      "eval_loss": 1.5133135318756104,
+      "eval_runtime": 89.6337,
+      "eval_samples_per_second": 6.426,
+      "eval_steps_per_second": 1.607,
+      "step": 220
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 7.349537037037037e-05,
+      "loss": 1.5165,
+      "step": 230
+    },
+    {
+      "epoch": 0.8,
+      "eval_accuracy": 0.4322916567325592,
+      "eval_loss": 1.4692301750183105,
+      "eval_runtime": 90.2666,
+      "eval_samples_per_second": 6.381,
+      "eval_steps_per_second": 1.595,
+      "step": 230
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 7.233796296296297e-05,
+      "loss": 1.4743,
+      "step": 240
+    },
+    {
+      "epoch": 0.83,
+      "eval_accuracy": 0.3975694477558136,
+      "eval_loss": 1.5509642362594604,
+      "eval_runtime": 90.8282,
+      "eval_samples_per_second": 6.342,
+      "eval_steps_per_second": 1.585,
+      "step": 240
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 7.118055555555556e-05,
+      "loss": 1.4903,
+      "step": 250
+    },
+    {
+      "epoch": 0.87,
+      "eval_accuracy": 0.4739583432674408,
+      "eval_loss": 1.3426711559295654,
+      "eval_runtime": 91.1121,
+      "eval_samples_per_second": 6.322,
+      "eval_steps_per_second": 1.58,
+      "step": 250
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.002314814814816e-05,
+      "loss": 1.2193,
+      "step": 260
+    },
+    {
+      "epoch": 0.9,
+      "eval_accuracy": 0.5329861044883728,
+      "eval_loss": 1.3175561428070068,
+      "eval_runtime": 89.5523,
+      "eval_samples_per_second": 6.432,
+      "eval_steps_per_second": 1.608,
+      "step": 260
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.886574074074074e-05,
+      "loss": 1.56,
+      "step": 270
+    },
+    {
+      "epoch": 0.94,
+      "eval_accuracy": 0.4635416567325592,
+      "eval_loss": 1.4102365970611572,
+      "eval_runtime": 89.6275,
+      "eval_samples_per_second": 6.427,
+      "eval_steps_per_second": 1.607,
+      "step": 270
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.770833333333334e-05,
+      "loss": 1.4563,
+      "step": 280
+    },
+    {
+      "epoch": 0.97,
+      "eval_accuracy": 0.5347222089767456,
+      "eval_loss": 1.291764736175537,
+      "eval_runtime": 89.5795,
+      "eval_samples_per_second": 6.43,
+      "eval_steps_per_second": 1.608,
+      "step": 280
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 6.655092592592593e-05,
+      "loss": 1.3766,
+      "step": 290
+    },
+    {
+      "epoch": 1.01,
+      "eval_accuracy": 0.4305555522441864,
+      "eval_loss": 1.5331988334655762,
+      "eval_runtime": 89.8777,
+      "eval_samples_per_second": 6.409,
+      "eval_steps_per_second": 1.602,
+      "step": 290
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 6.539351851851853e-05,
+      "loss": 1.218,
+      "step": 300
+    },
+    {
+      "epoch": 1.04,
+      "eval_accuracy": 0.4861111044883728,
+      "eval_loss": 1.317336916923523,
+      "eval_runtime": 91.03,
+      "eval_samples_per_second": 6.328,
+      "eval_steps_per_second": 1.582,
+      "step": 300
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 6.423611111111112e-05,
+      "loss": 1.3211,
+      "step": 310
+    },
+    {
+      "epoch": 1.08,
+      "eval_accuracy": 0.5034722089767456,
+      "eval_loss": 1.263780951499939,
+      "eval_runtime": 90.5509,
+      "eval_samples_per_second": 6.361,
+      "eval_steps_per_second": 1.59,
+      "step": 310
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 6.307870370370372e-05,
+      "loss": 1.1933,
+      "step": 320
+    },
+    {
+      "epoch": 1.11,
+      "eval_accuracy": 0.4982638955116272,
+      "eval_loss": 1.3839294910430908,
+      "eval_runtime": 90.7629,
+      "eval_samples_per_second": 6.346,
+      "eval_steps_per_second": 1.587,
+      "step": 320
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 6.192129629629629e-05,
+      "loss": 1.2803,
+      "step": 330
+    },
+    {
+      "epoch": 1.15,
+      "eval_accuracy": 0.5121527910232544,
+      "eval_loss": 1.2145192623138428,
+      "eval_runtime": 90.0175,
+      "eval_samples_per_second": 6.399,
+      "eval_steps_per_second": 1.6,
+      "step": 330
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 6.076388888888889e-05,
+      "loss": 1.1112,
+      "step": 340
+    },
+    {
+      "epoch": 1.18,
+      "eval_accuracy": 0.5711805820465088,
+      "eval_loss": 1.1930283308029175,
+      "eval_runtime": 89.5796,
+      "eval_samples_per_second": 6.43,
+      "eval_steps_per_second": 1.608,
+      "step": 340
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 5.960648148148148e-05,
+      "loss": 1.0907,
+      "step": 350
+    },
+    {
+      "epoch": 1.22,
+      "eval_accuracy": 0.5815972089767456,
+      "eval_loss": 1.145843505859375,
+      "eval_runtime": 90.3053,
+      "eval_samples_per_second": 6.378,
+      "eval_steps_per_second": 1.595,
+      "step": 350
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 5.844907407407407e-05,
+      "loss": 1.013,
+      "step": 360
+    },
+    {
+      "epoch": 1.25,
+      "eval_accuracy": 0.6180555820465088,
+      "eval_loss": 1.0559839010238647,
+      "eval_runtime": 89.8917,
+      "eval_samples_per_second": 6.408,
+      "eval_steps_per_second": 1.602,
+      "step": 360
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 5.7291666666666666e-05,
+      "loss": 0.9308,
+      "step": 370
+    },
+    {
+      "epoch": 1.28,
+      "eval_accuracy": 0.6163194179534912,
+      "eval_loss": 1.0486806631088257,
+      "eval_runtime": 90.7378,
+      "eval_samples_per_second": 6.348,
+      "eval_steps_per_second": 1.587,
+      "step": 370
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.613425925925926e-05,
+      "loss": 1.3225,
+      "step": 380
+    },
+    {
+      "epoch": 1.32,
+      "eval_accuracy": 0.6145833134651184,
+      "eval_loss": 1.050321340560913,
+      "eval_runtime": 89.4251,
+      "eval_samples_per_second": 6.441,
+      "eval_steps_per_second": 1.61,
+      "step": 380
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.497685185185185e-05,
+      "loss": 1.0774,
+      "step": 390
+    },
+    {
+      "epoch": 1.35,
+      "eval_accuracy": 0.5434027910232544,
+      "eval_loss": 1.25161612033844,
+      "eval_runtime": 89.4532,
+      "eval_samples_per_second": 6.439,
+      "eval_steps_per_second": 1.61,
+      "step": 390
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 5.3819444444444444e-05,
+      "loss": 1.2251,
+      "step": 400
+    },
+    {
+      "epoch": 1.39,
+      "eval_accuracy": 0.609375,
+      "eval_loss": 1.076072335243225,
+      "eval_runtime": 90.2082,
+      "eval_samples_per_second": 6.385,
+      "eval_steps_per_second": 1.596,
+      "step": 400
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 5.266203703703704e-05,
+      "loss": 0.9848,
+      "step": 410
+    },
+    {
+      "epoch": 1.42,
+      "eval_accuracy": 0.6440972089767456,
+      "eval_loss": 1.0271832942962646,
+      "eval_runtime": 89.9601,
+      "eval_samples_per_second": 6.403,
+      "eval_steps_per_second": 1.601,
+      "step": 410
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.150462962962963e-05,
+      "loss": 0.9913,
+      "step": 420
+    },
+    {
+      "epoch": 1.46,
+      "eval_accuracy": 0.6041666865348816,
+      "eval_loss": 1.0503506660461426,
+      "eval_runtime": 89.8786,
+      "eval_samples_per_second": 6.409,
+      "eval_steps_per_second": 1.602,
+      "step": 420
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 5.034722222222222e-05,
+      "loss": 0.9081,
+      "step": 430
+    },
+    {
+      "epoch": 1.49,
+      "eval_accuracy": 0.6666666865348816,
+      "eval_loss": 0.9094821214675903,
+      "eval_runtime": 91.6907,
+      "eval_samples_per_second": 6.282,
+      "eval_steps_per_second": 1.57,
+      "step": 430
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.9189814814814815e-05,
+      "loss": 0.8339,
+      "step": 440
+    },
+    {
+      "epoch": 1.53,
+      "eval_accuracy": 0.6631944179534912,
+      "eval_loss": 0.9030921459197998,
+      "eval_runtime": 90.547,
+      "eval_samples_per_second": 6.361,
+      "eval_steps_per_second": 1.59,
+      "step": 440
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.803240740740741e-05,
+      "loss": 0.8893,
+      "step": 450
+    },
+    {
+      "epoch": 1.56,
+      "eval_accuracy": 0.6423611044883728,
+      "eval_loss": 0.9375382661819458,
+      "eval_runtime": 90.2281,
+      "eval_samples_per_second": 6.384,
+      "eval_steps_per_second": 1.596,
+      "step": 450
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.6875e-05,
+      "loss": 0.9362,
+      "step": 460
+    },
+    {
+      "epoch": 1.6,
+      "eval_accuracy": 0.6197916865348816,
+      "eval_loss": 0.9755175113677979,
+      "eval_runtime": 91.0163,
+      "eval_samples_per_second": 6.329,
+      "eval_steps_per_second": 1.582,
+      "step": 460
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.5717592592592594e-05,
+      "loss": 0.835,
+      "step": 470
+    },
+    {
+      "epoch": 1.63,
+      "eval_accuracy": 0.6545138955116272,
+      "eval_loss": 0.9399816989898682,
+      "eval_runtime": 89.8555,
+      "eval_samples_per_second": 6.41,
+      "eval_steps_per_second": 1.603,
+      "step": 470
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.456018518518519e-05,
+      "loss": 0.6733,
+      "step": 480
+    },
+    {
+      "epoch": 1.67,
+      "eval_accuracy": 0.6927083134651184,
+      "eval_loss": 0.8480438590049744,
+      "eval_runtime": 89.8841,
+      "eval_samples_per_second": 6.408,
+      "eval_steps_per_second": 1.602,
+      "step": 480
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.340277777777778e-05,
+      "loss": 1.0115,
+      "step": 490
+    },
+    {
+      "epoch": 1.7,
+      "eval_accuracy": 0.6840277910232544,
+      "eval_loss": 0.8332173824310303,
+      "eval_runtime": 91.0011,
+      "eval_samples_per_second": 6.33,
+      "eval_steps_per_second": 1.582,
+      "step": 490
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.224537037037037e-05,
+      "loss": 0.7473,
+      "step": 500
+    },
+    {
+      "epoch": 1.74,
+      "eval_accuracy": 0.6475694179534912,
+      "eval_loss": 0.9618370532989502,
+      "eval_runtime": 90.2515,
+      "eval_samples_per_second": 6.382,
+      "eval_steps_per_second": 1.596,
+      "step": 500
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.1087962962962965e-05,
+      "loss": 0.8355,
+      "step": 510
+    },
+    {
+      "epoch": 1.77,
+      "eval_accuracy": 0.6840277910232544,
+      "eval_loss": 0.8845413327217102,
+      "eval_runtime": 89.7252,
+      "eval_samples_per_second": 6.42,
+      "eval_steps_per_second": 1.605,
+      "step": 510
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.993055555555556e-05,
+      "loss": 0.8487,
+      "step": 520
+    },
+    {
+      "epoch": 1.81,
+      "eval_accuracy": 0.6875,
+      "eval_loss": 0.8297374844551086,
+      "eval_runtime": 91.9685,
+      "eval_samples_per_second": 6.263,
+      "eval_steps_per_second": 1.566,
+      "step": 520
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.877314814814815e-05,
+      "loss": 0.6038,
+      "step": 530
+    },
+    {
+      "epoch": 1.84,
+      "eval_accuracy": 0.6493055820465088,
+      "eval_loss": 0.9539130330085754,
+      "eval_runtime": 90.0856,
+      "eval_samples_per_second": 6.394,
+      "eval_steps_per_second": 1.598,
+      "step": 530
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.7615740740740744e-05,
+      "loss": 0.75,
+      "step": 540
+    },
+    {
+      "epoch": 1.88,
+      "eval_accuracy": 0.6857638955116272,
+      "eval_loss": 0.8455307483673096,
+      "eval_runtime": 89.5522,
+      "eval_samples_per_second": 6.432,
+      "eval_steps_per_second": 1.608,
+      "step": 540
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.6458333333333336e-05,
+      "loss": 0.8561,
+      "step": 550
+    },
+    {
+      "epoch": 1.91,
+      "eval_accuracy": 0.7013888955116272,
+      "eval_loss": 0.7813519239425659,
+      "eval_runtime": 90.1129,
+      "eval_samples_per_second": 6.392,
+      "eval_steps_per_second": 1.598,
+      "step": 550
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.530092592592593e-05,
+      "loss": 0.7552,
+      "step": 560
+    },
+    {
+      "epoch": 1.94,
+      "eval_accuracy": 0.6822916865348816,
+      "eval_loss": 0.8651251196861267,
+      "eval_runtime": 89.9146,
+      "eval_samples_per_second": 6.406,
+      "eval_steps_per_second": 1.602,
+      "step": 560
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.414351851851852e-05,
+      "loss": 0.6972,
+      "step": 570
+    },
+    {
+      "epoch": 1.98,
+      "eval_accuracy": 0.71875,
+      "eval_loss": 0.7325252890586853,
+      "eval_runtime": 90.7375,
+      "eval_samples_per_second": 6.348,
+      "eval_steps_per_second": 1.587,
+      "step": 570
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 3.2986111111111115e-05,
+      "loss": 0.7483,
+      "step": 580
+    },
+    {
+      "epoch": 2.01,
+      "eval_accuracy": 0.7690972089767456,
+      "eval_loss": 0.6722133159637451,
+      "eval_runtime": 90.3148,
+      "eval_samples_per_second": 6.378,
+      "eval_steps_per_second": 1.594,
+      "step": 580
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 3.182870370370371e-05,
+      "loss": 0.5419,
+      "step": 590
+    },
+    {
+      "epoch": 2.05,
+      "eval_accuracy": 0.7326388955116272,
+      "eval_loss": 0.7046216130256653,
+      "eval_runtime": 92.1898,
+      "eval_samples_per_second": 6.248,
+      "eval_steps_per_second": 1.562,
+      "step": 590
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.06712962962963e-05,
+      "loss": 0.5203,
+      "step": 600
+    },
+    {
+      "epoch": 2.08,
+      "eval_accuracy": 0.7326388955116272,
+      "eval_loss": 0.7062063813209534,
+      "eval_runtime": 90.2661,
+      "eval_samples_per_second": 6.381,
+      "eval_steps_per_second": 1.595,
+      "step": 600
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.951388888888889e-05,
+      "loss": 0.5235,
+      "step": 610
+    },
+    {
+      "epoch": 2.12,
+      "eval_accuracy": 0.7534722089767456,
+      "eval_loss": 0.6795992255210876,
+      "eval_runtime": 91.0148,
+      "eval_samples_per_second": 6.329,
+      "eval_steps_per_second": 1.582,
+      "step": 610
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.8356481481481483e-05,
+      "loss": 0.514,
+      "step": 620
+    },
+    {
+      "epoch": 2.15,
+      "eval_accuracy": 0.7204861044883728,
+      "eval_loss": 0.746653139591217,
+      "eval_runtime": 89.8044,
+      "eval_samples_per_second": 6.414,
+      "eval_steps_per_second": 1.603,
+      "step": 620
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.7199074074074076e-05,
+      "loss": 0.5402,
+      "step": 630
+    },
+    {
+      "epoch": 2.19,
+      "eval_accuracy": 0.7447916865348816,
+      "eval_loss": 0.716274619102478,
+      "eval_runtime": 90.8833,
+      "eval_samples_per_second": 6.338,
+      "eval_steps_per_second": 1.584,
+      "step": 630
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.604166666666667e-05,
+      "loss": 0.7235,
+      "step": 640
+    },
+    {
+      "epoch": 2.22,
+      "eval_accuracy": 0.7222222089767456,
+      "eval_loss": 0.7545790076255798,
+      "eval_runtime": 89.6633,
+      "eval_samples_per_second": 6.424,
+      "eval_steps_per_second": 1.606,
+      "step": 640
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 2.488425925925926e-05,
+      "loss": 0.551,
+      "step": 650
+    },
+    {
+      "epoch": 2.26,
+      "eval_accuracy": 0.7534722089767456,
+      "eval_loss": 0.6994116902351379,
+      "eval_runtime": 89.8085,
+      "eval_samples_per_second": 6.414,
+      "eval_steps_per_second": 1.603,
+      "step": 650
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.3726851851851854e-05,
+      "loss": 0.5769,
+      "step": 660
+    },
+    {
+      "epoch": 2.29,
+      "eval_accuracy": 0.7534722089767456,
+      "eval_loss": 0.7151244282722473,
+      "eval_runtime": 90.9508,
+      "eval_samples_per_second": 6.333,
+      "eval_steps_per_second": 1.583,
+      "step": 660
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 2.2569444444444447e-05,
+      "loss": 0.5501,
+      "step": 670
+    },
+    {
+      "epoch": 2.33,
+      "eval_accuracy": 0.7604166865348816,
+      "eval_loss": 0.695513129234314,
+      "eval_runtime": 90.6092,
+      "eval_samples_per_second": 6.357,
+      "eval_steps_per_second": 1.589,
+      "step": 670
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 2.141203703703704e-05,
+      "loss": 0.5416,
+      "step": 680
+    },
+    {
+      "epoch": 2.36,
+      "eval_accuracy": 0.7725694179534912,
+      "eval_loss": 0.6533116102218628,
+      "eval_runtime": 89.82,
+      "eval_samples_per_second": 6.413,
+      "eval_steps_per_second": 1.603,
+      "step": 680
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.0254629629629632e-05,
+      "loss": 0.5452,
+      "step": 690
+    },
+    {
+      "epoch": 2.4,
+      "eval_accuracy": 0.7777777910232544,
+      "eval_loss": 0.6232606172561646,
+      "eval_runtime": 90.4116,
+      "eval_samples_per_second": 6.371,
+      "eval_steps_per_second": 1.593,
+      "step": 690
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.9097222222222222e-05,
+      "loss": 0.8518,
+      "step": 700
+    },
+    {
+      "epoch": 2.43,
+      "eval_accuracy": 0.7777777910232544,
+      "eval_loss": 0.6136298179626465,
+      "eval_runtime": 90.8648,
+      "eval_samples_per_second": 6.339,
+      "eval_steps_per_second": 1.585,
+      "step": 700
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.7939814814814815e-05,
+      "loss": 0.3372,
+      "step": 710
+    },
+    {
+      "epoch": 2.47,
+      "eval_accuracy": 0.7986111044883728,
+      "eval_loss": 0.5700623393058777,
+      "eval_runtime": 90.1438,
+      "eval_samples_per_second": 6.39,
+      "eval_steps_per_second": 1.597,
+      "step": 710
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.6782407407407408e-05,
+      "loss": 0.4488,
+      "step": 720
+    },
+    {
+      "epoch": 2.5,
+      "eval_accuracy": 0.7847222089767456,
+      "eval_loss": 0.5789040327072144,
+      "eval_runtime": 90.7633,
+      "eval_samples_per_second": 6.346,
+      "eval_steps_per_second": 1.587,
+      "step": 720
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.5625e-05,
+      "loss": 0.3977,
+      "step": 730
+    },
+    {
+      "epoch": 2.53,
+      "eval_accuracy": 0.7829861044883728,
+      "eval_loss": 0.5748720169067383,
+      "eval_runtime": 89.6735,
+      "eval_samples_per_second": 6.423,
+      "eval_steps_per_second": 1.606,
+      "step": 730
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.4467592592592593e-05,
+      "loss": 0.3625,
+      "step": 740
+    },
+    {
+      "epoch": 2.57,
+      "eval_accuracy": 0.8003472089767456,
+      "eval_loss": 0.5525452494621277,
+      "eval_runtime": 90.5733,
+      "eval_samples_per_second": 6.359,
+      "eval_steps_per_second": 1.59,
+      "step": 740
     }
   ],
-  "max_steps": 100,
-  "num_train_epochs": 1,
-  "total_flos": 1.4748389432630784e+17,
+  "max_steps": 864,
+  "num_train_epochs": 3,
+  "total_flos": 2.187634079192584e+18,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:70a15c44c864138c37d3beb805073febdbcd67d33560564bdcb6bc07ab6c0653
+oid sha256:2f6e10d1da3f25a338789486f1c3c671befb6ab6edeb2ae72a2760c485d4e12b
 size 3439