Training in progress, step 500, checkpoint

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:438da24f300823200d0f4f1d495e9c6e8c220ca1df307a75c80aaa41bcefba4e
+oid sha256:701e6d0bbf056ae41ebcd2ddad1bad3601d6ab1f0e5dd5ac0c2525d0668aa1cc
 size 410636248

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dcc62cb139e4056ac7ba199ccdf3b32ce7584d0c69fbaa087876feaee35c8e68
+oid sha256:ce59802aba0b143b17031171bed3296a63bf09e719029281cb6fa94a848a4b66
 size 821393658

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cb0f71048b41c58f8b88689478609e70f109b40527076516f1d053575248b926
+oid sha256:a140ab4fb1cc3180397e45a099073097a2e2e1e4a8db652ca5718334717ca6cf
 size 14244

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c3ea97f3d2e35b33593e80f0c73d6f3560648e3a826ab4155c72d23d65cd009d
+oid sha256:e3ab3fae6bcbaddd22accfbc804c2c716cc71d2097adfaaa313faeb8455aa736
 size 1064

last-checkpoint/trainer_state.json

@@ -1,812 +1,412 @@
 {
-  "best_metric": 0.10818465054035187,
-  "best_model_checkpoint": "./results/checkpoint-1000",
-  "epoch": 4.8076923076923075,
+  "best_metric": 0.16280439496040344,
+  "best_model_checkpoint": "./results/checkpoint-500",
+  "epoch": 2.4038461538461537,
   "eval_steps": 20,
-  "global_step": 1000,
+  "global_step": 500,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.09615384615384616,
-      "grad_norm": 2.879941701889038,
+      "grad_norm": 1.8630799055099487,
       "learning_rate": 1.9615384615384617e-05,
-      "loss": 0.5169,
+      "loss": 0.4117,
       "step": 20
     },
     {
       "epoch": 0.09615384615384616,
-      "eval_accuracy": 0.8064516129032258,
-      "eval_loss": 0.43645337224006653,
-      "eval_runtime": 2.8523,
-      "eval_samples_per_second": 32.606,
-      "eval_steps_per_second": 8.414,
+      "eval_accuracy": 0.8387096774193549,
+      "eval_loss": 0.29503169655799866,
+      "eval_runtime": 2.9188,
+      "eval_samples_per_second": 31.863,
+      "eval_steps_per_second": 8.223,
       "step": 20
     },
     {
       "epoch": 0.19230769230769232,
-      "grad_norm": 0.7589161396026611,
+      "grad_norm": 9.186605453491211,
       "learning_rate": 1.923076923076923e-05,
-      "loss": 0.3487,
+      "loss": 0.3727,
       "step": 40
     },
     {
       "epoch": 0.19230769230769232,
-      "eval_accuracy": 0.8602150537634409,
-      "eval_loss": 0.32320985198020935,
-      "eval_runtime": 2.9393,
-      "eval_samples_per_second": 31.641,
-      "eval_steps_per_second": 8.165,
+      "eval_accuracy": 0.8709677419354839,
+      "eval_loss": 0.20473997294902802,
+      "eval_runtime": 2.8638,
+      "eval_samples_per_second": 32.475,
+      "eval_steps_per_second": 8.381,
       "step": 40
     },
     {
       "epoch": 0.28846153846153844,
-      "grad_norm": 22.88661766052246,
+      "grad_norm": 8.3997802734375,
       "learning_rate": 1.8846153846153846e-05,
-      "loss": 0.1785,
+      "loss": 0.2863,
       "step": 60
     },
     {
       "epoch": 0.28846153846153844,
-      "eval_accuracy": 0.946236559139785,
-      "eval_loss": 0.1469801366329193,
-      "eval_runtime": 3.241,
-      "eval_samples_per_second": 28.695,
-      "eval_steps_per_second": 7.405,
+      "eval_accuracy": 0.956989247311828,
+      "eval_loss": 0.2754175662994385,
+      "eval_runtime": 2.7512,
+      "eval_samples_per_second": 33.803,
+      "eval_steps_per_second": 8.723,
       "step": 60
     },
     {
       "epoch": 0.38461538461538464,
-      "grad_norm": 0.04444374889135361,
+      "grad_norm": 0.0750838965177536,
       "learning_rate": 1.8461538461538465e-05,
-      "loss": 0.192,
+      "loss": 0.2698,
       "step": 80
     },
     {
       "epoch": 0.38461538461538464,
-      "eval_accuracy": 0.956989247311828,
-      "eval_loss": 0.12639252841472626,
-      "eval_runtime": 2.8312,
-      "eval_samples_per_second": 32.849,
-      "eval_steps_per_second": 8.477,
+      "eval_accuracy": 0.946236559139785,
+      "eval_loss": 0.16919000446796417,
+      "eval_runtime": 2.6759,
+      "eval_samples_per_second": 34.754,
+      "eval_steps_per_second": 8.969,
       "step": 80
     },
     {
       "epoch": 0.4807692307692308,
-      "grad_norm": 1.9381160736083984,
+      "grad_norm": 0.0878710225224495,
       "learning_rate": 1.807692307692308e-05,
-      "loss": 0.0329,
+      "loss": 0.1232,
       "step": 100
     },
     {
       "epoch": 0.4807692307692308,
-      "eval_accuracy": 0.956989247311828,
-      "eval_loss": 0.18568147718906403,
-      "eval_runtime": 2.7388,
-      "eval_samples_per_second": 33.956,
-      "eval_steps_per_second": 8.763,
+      "eval_accuracy": 0.946236559139785,
+      "eval_loss": 0.16618885099887848,
+      "eval_runtime": 2.6422,
+      "eval_samples_per_second": 35.197,
+      "eval_steps_per_second": 9.083,
       "step": 100
     },
     {
       "epoch": 0.5769230769230769,
-      "grad_norm": 0.0944330021739006,
+      "grad_norm": 0.038383062928915024,
       "learning_rate": 1.7692307692307694e-05,
-      "loss": 0.1432,
+      "loss": 0.0806,
       "step": 120
     },
     {
       "epoch": 0.5769230769230769,
-      "eval_accuracy": 0.9247311827956989,
-      "eval_loss": 0.2023310661315918,
-      "eval_runtime": 2.6839,
-      "eval_samples_per_second": 34.651,
-      "eval_steps_per_second": 8.942,
+      "eval_accuracy": 0.946236559139785,
+      "eval_loss": 0.15975366532802582,
+      "eval_runtime": 2.6652,
+      "eval_samples_per_second": 34.894,
+      "eval_steps_per_second": 9.005,
       "step": 120
     },
     {
       "epoch": 0.6730769230769231,
-      "grad_norm": 36.735008239746094,
+      "grad_norm": 0.022622136399149895,
       "learning_rate": 1.730769230769231e-05,
-      "loss": 0.3458,
+      "loss": 0.2571,
       "step": 140
     },
     {
       "epoch": 0.6730769230769231,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.12693293392658234,
-      "eval_runtime": 2.6852,
-      "eval_samples_per_second": 34.634,
-      "eval_steps_per_second": 8.938,
+      "eval_accuracy": 0.978494623655914,
+      "eval_loss": 0.0726698786020279,
+      "eval_runtime": 2.6685,
+      "eval_samples_per_second": 34.85,
+      "eval_steps_per_second": 8.994,
       "step": 140
     },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 13.383593559265137,
+      "grad_norm": 9.369976043701172,
       "learning_rate": 1.6923076923076924e-05,
-      "loss": 0.234,
+      "loss": 0.1374,
       "step": 160
     },
     {
       "epoch": 0.7692307692307693,
-      "eval_accuracy": 0.946236559139785,
-      "eval_loss": 0.15099573135375977,
-      "eval_runtime": 2.7092,
-      "eval_samples_per_second": 34.327,
-      "eval_steps_per_second": 8.859,
+      "eval_accuracy": 0.978494623655914,
+      "eval_loss": 0.0701598972082138,
+      "eval_runtime": 2.6859,
+      "eval_samples_per_second": 34.626,
+      "eval_steps_per_second": 8.936,
       "step": 160
     },
     {
       "epoch": 0.8653846153846154,
-      "grad_norm": 0.04881865903735161,
+      "grad_norm": 7.6264119148254395,
       "learning_rate": 1.653846153846154e-05,
-      "loss": 0.179,
+      "loss": 0.165,
       "step": 180
     },
     {
       "epoch": 0.8653846153846154,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.11019934713840485,
-      "eval_runtime": 2.7609,
-      "eval_samples_per_second": 33.685,
-      "eval_steps_per_second": 8.693,
+      "eval_accuracy": 0.956989247311828,
+      "eval_loss": 0.08651256561279297,
+      "eval_runtime": 2.7028,
+      "eval_samples_per_second": 34.409,
+      "eval_steps_per_second": 8.88,
       "step": 180
     },
     {
       "epoch": 0.9615384615384616,
-      "grad_norm": 0.1799299567937851,
+      "grad_norm": 0.03415932506322861,
       "learning_rate": 1.6153846153846154e-05,
-      "loss": 0.1873,
+      "loss": 0.528,
       "step": 200
     },
     {
       "epoch": 0.9615384615384616,
-      "eval_accuracy": 0.956989247311828,
-      "eval_loss": 0.11402961611747742,
-      "eval_runtime": 2.7766,
-      "eval_samples_per_second": 33.494,
-      "eval_steps_per_second": 8.644,
+      "eval_accuracy": 0.967741935483871,
+      "eval_loss": 0.12613914906978607,
+      "eval_runtime": 2.7168,
+      "eval_samples_per_second": 34.231,
+      "eval_steps_per_second": 8.834,
       "step": 200
     },
     {
       "epoch": 1.0576923076923077,
-      "grad_norm": 0.45117759704589844,
+      "grad_norm": 0.17008128762245178,
       "learning_rate": 1.576923076923077e-05,
-      "loss": 0.1519,
+      "loss": 0.1023,
       "step": 220
     },
     {
       "epoch": 1.0576923076923077,
-      "eval_accuracy": 0.946236559139785,
-      "eval_loss": 0.2102302461862564,
-      "eval_runtime": 2.7619,
-      "eval_samples_per_second": 33.672,
-      "eval_steps_per_second": 8.69,
+      "eval_accuracy": 0.9139784946236559,
+      "eval_loss": 0.37878894805908203,
+      "eval_runtime": 2.7414,
+      "eval_samples_per_second": 33.924,
+      "eval_steps_per_second": 8.755,
       "step": 220
     },
     {
       "epoch": 1.1538461538461537,
-      "grad_norm": 55.769256591796875,
+      "grad_norm": 0.010861682705581188,
       "learning_rate": 1.5384615384615387e-05,
-      "loss": 0.0453,
+      "loss": 0.0278,
       "step": 240
     },
     {
       "epoch": 1.1538461538461537,
-      "eval_accuracy": 0.9247311827956989,
-      "eval_loss": 0.31499457359313965,
-      "eval_runtime": 2.7752,
-      "eval_samples_per_second": 33.511,
-      "eval_steps_per_second": 8.648,
+      "eval_accuracy": 0.946236559139785,
+      "eval_loss": 0.1893940567970276,
+      "eval_runtime": 2.7091,
+      "eval_samples_per_second": 34.329,
+      "eval_steps_per_second": 8.859,
       "step": 240
     },
     {
       "epoch": 1.25,
-      "grad_norm": 0.028473777696490288,
+      "grad_norm": 0.017718007788062096,
       "learning_rate": 1.5000000000000002e-05,
-      "loss": 0.0341,
+      "loss": 0.1642,
       "step": 260
     },
     {
       "epoch": 1.25,
-      "eval_accuracy": 0.956989247311828,
-      "eval_loss": 0.1401311606168747,
-      "eval_runtime": 2.7614,
-      "eval_samples_per_second": 33.678,
-      "eval_steps_per_second": 8.691,
+      "eval_accuracy": 0.9139784946236559,
+      "eval_loss": 0.3861010670661926,
+      "eval_runtime": 2.6955,
+      "eval_samples_per_second": 34.503,
+      "eval_steps_per_second": 8.904,
       "step": 260
     },
     {
       "epoch": 1.3461538461538463,
-      "grad_norm": 0.04881782829761505,
+      "grad_norm": 0.021516259759664536,
       "learning_rate": 1.4615384615384615e-05,
-      "loss": 0.0316,
+      "loss": 0.1108,
       "step": 280
     },
     {
       "epoch": 1.3461538461538463,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.16341440379619598,
-      "eval_runtime": 2.7702,
-      "eval_samples_per_second": 33.572,
-      "eval_steps_per_second": 8.664,
+      "eval_accuracy": 0.9354838709677419,
+      "eval_loss": 0.20804111659526825,
+      "eval_runtime": 2.9167,
+      "eval_samples_per_second": 31.886,
+      "eval_steps_per_second": 8.229,
       "step": 280
     },
     {
       "epoch": 1.4423076923076923,
-      "grad_norm": 0.040301088243722916,
+      "grad_norm": 0.013441790826618671,
       "learning_rate": 1.4230769230769232e-05,
-      "loss": 0.1082,
+      "loss": 0.1514,
       "step": 300
     },
     {
       "epoch": 1.4423076923076923,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.10686944425106049,
-      "eval_runtime": 2.7502,
-      "eval_samples_per_second": 33.816,
-      "eval_steps_per_second": 8.727,
+      "eval_accuracy": 0.967741935483871,
+      "eval_loss": 0.15365849435329437,
+      "eval_runtime": 2.8225,
+      "eval_samples_per_second": 32.95,
+      "eval_steps_per_second": 8.503,
       "step": 300
     },
     {
       "epoch": 1.5384615384615383,
-      "grad_norm": 27.52263641357422,
+      "grad_norm": 0.02140916883945465,
       "learning_rate": 1.3846153846153847e-05,
-      "loss": 0.0954,
+      "loss": 0.0025,
       "step": 320
     },
     {
       "epoch": 1.5384615384615383,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.08186855167150497,
-      "eval_runtime": 2.7311,
-      "eval_samples_per_second": 34.052,
-      "eval_steps_per_second": 8.788,
+      "eval_accuracy": 0.946236559139785,
+      "eval_loss": 0.30473944544792175,
+      "eval_runtime": 2.6926,
+      "eval_samples_per_second": 34.539,
+      "eval_steps_per_second": 8.913,
       "step": 320
     },
     {
       "epoch": 1.6346153846153846,
-      "grad_norm": 0.013332587666809559,
+      "grad_norm": 0.011240499094128609,
       "learning_rate": 1.3461538461538463e-05,
-      "loss": 0.0472,
+      "loss": 0.1124,
       "step": 340
     },
     {
       "epoch": 1.6346153846153846,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.1686173677444458,
-      "eval_runtime": 2.733,
-      "eval_samples_per_second": 34.029,
-      "eval_steps_per_second": 8.782,
+      "eval_accuracy": 0.9354838709677419,
+      "eval_loss": 0.24811075627803802,
+      "eval_runtime": 2.6909,
+      "eval_samples_per_second": 34.561,
+      "eval_steps_per_second": 8.919,
       "step": 340
     },
     {
       "epoch": 1.7307692307692308,
-      "grad_norm": 0.05663549154996872,
+      "grad_norm": 0.009658828377723694,
       "learning_rate": 1.3076923076923078e-05,
-      "loss": 0.1563,
+      "loss": 0.3081,
       "step": 360
     },
     {
       "epoch": 1.7307692307692308,
       "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.037865716964006424,
-      "eval_runtime": 2.7489,
-      "eval_samples_per_second": 33.832,
-      "eval_steps_per_second": 8.731,
+      "eval_loss": 0.1271650344133377,
+      "eval_runtime": 2.6952,
+      "eval_samples_per_second": 34.505,
+      "eval_steps_per_second": 8.905,
       "step": 360
     },
     {
       "epoch": 1.8269230769230769,
-      "grad_norm": 8.364692687988281,
+      "grad_norm": 48.972293853759766,
       "learning_rate": 1.2692307692307693e-05,
-      "loss": 0.1812,
+      "loss": 0.0764,
       "step": 380
     },
     {
       "epoch": 1.8269230769230769,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.12184549123048782,
-      "eval_runtime": 2.7376,
-      "eval_samples_per_second": 33.971,
-      "eval_steps_per_second": 8.767,
+      "eval_accuracy": 0.989247311827957,
+      "eval_loss": 0.0935346931219101,
+      "eval_runtime": 2.7195,
+      "eval_samples_per_second": 34.198,
+      "eval_steps_per_second": 8.825,
       "step": 380
     },
     {
       "epoch": 1.9230769230769231,
-      "grad_norm": 0.173508420586586,
+      "grad_norm": 0.01087904442101717,
       "learning_rate": 1.230769230769231e-05,
-      "loss": 0.1276,
+      "loss": 0.0923,
       "step": 400
     },
     {
       "epoch": 1.9230769230769231,
-      "eval_accuracy": 0.989247311827957,
-      "eval_loss": 0.07845792919397354,
-      "eval_runtime": 2.7605,
-      "eval_samples_per_second": 33.689,
-      "eval_steps_per_second": 8.694,
+      "eval_accuracy": 0.9354838709677419,
+      "eval_loss": 0.23266802728176117,
+      "eval_runtime": 2.7195,
+      "eval_samples_per_second": 34.197,
+      "eval_steps_per_second": 8.825,
       "step": 400
     },
     {
       "epoch": 2.019230769230769,
-      "grad_norm": 0.009747051633894444,
+      "grad_norm": 19.043825149536133,
       "learning_rate": 1.1923076923076925e-05,
-      "loss": 0.0772,
+      "loss": 0.137,
       "step": 420
     },
     {
       "epoch": 2.019230769230769,
-      "eval_accuracy": 0.989247311827957,
-      "eval_loss": 0.07875293493270874,
-      "eval_runtime": 2.7654,
-      "eval_samples_per_second": 33.63,
-      "eval_steps_per_second": 8.679,
+      "eval_accuracy": 0.956989247311828,
+      "eval_loss": 0.16817408800125122,
+      "eval_runtime": 2.7341,
+      "eval_samples_per_second": 34.015,
+      "eval_steps_per_second": 8.778,
       "step": 420
     },
     {
       "epoch": 2.1153846153846154,
-      "grad_norm": 0.017300551757216454,
+      "grad_norm": 0.036368228495121,
       "learning_rate": 1.1538461538461538e-05,
-      "loss": 0.0022,
+      "loss": 0.0796,
       "step": 440
     },
     {
       "epoch": 2.1153846153846154,
       "eval_accuracy": 0.956989247311828,
-      "eval_loss": 0.102827288210392,
-      "eval_runtime": 2.769,
-      "eval_samples_per_second": 33.587,
-      "eval_steps_per_second": 8.667,
+      "eval_loss": 0.16539430618286133,
+      "eval_runtime": 2.7265,
+      "eval_samples_per_second": 34.11,
+      "eval_steps_per_second": 8.803,
       "step": 440
     },
     {
       "epoch": 2.2115384615384617,
-      "grad_norm": 0.019045885652303696,
+      "grad_norm": 0.03690864145755768,
       "learning_rate": 1.1153846153846154e-05,
-      "loss": 0.0011,
+      "loss": 0.0089,
       "step": 460
     },
     {
       "epoch": 2.2115384615384617,
       "eval_accuracy": 0.956989247311828,
-      "eval_loss": 0.15620151162147522,
-      "eval_runtime": 2.7371,
-      "eval_samples_per_second": 33.977,
-      "eval_steps_per_second": 8.768,
+      "eval_loss": 0.2199721783399582,
+      "eval_runtime": 2.7071,
+      "eval_samples_per_second": 34.354,
+      "eval_steps_per_second": 8.865,
       "step": 460
     },
     {
       "epoch": 2.3076923076923075,
-      "grad_norm": 0.005280839279294014,
+      "grad_norm": 0.011880586855113506,
       "learning_rate": 1.076923076923077e-05,
-      "loss": 0.076,
+      "loss": 0.1514,
       "step": 480
     },
     {
       "epoch": 2.3076923076923075,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.12186174839735031,
-      "eval_runtime": 2.7523,
-      "eval_samples_per_second": 33.789,
-      "eval_steps_per_second": 8.72,
+      "eval_accuracy": 0.956989247311828,
+      "eval_loss": 0.20590856671333313,
+      "eval_runtime": 2.697,
+      "eval_samples_per_second": 34.482,
+      "eval_steps_per_second": 8.899,
       "step": 480
     },
     {
       "epoch": 2.4038461538461537,
-      "grad_norm": 0.01168849878013134,
+      "grad_norm": 0.008282246068120003,
       "learning_rate": 1.0384615384615386e-05,
-      "loss": 0.0012,
+      "loss": 0.0044,
       "step": 500
     },
     {
       "epoch": 2.4038461538461537,
       "eval_accuracy": 0.956989247311828,
-      "eval_loss": 0.11589507758617401,
-      "eval_runtime": 2.7518,
-      "eval_samples_per_second": 33.796,
-      "eval_steps_per_second": 8.721,
+      "eval_loss": 0.16280439496040344,
+      "eval_runtime": 2.7133,
+      "eval_samples_per_second": 34.275,
+      "eval_steps_per_second": 8.845,
       "step": 500
-    },
-    {
-      "epoch": 2.5,
-      "grad_norm": 0.00502841966226697,
-      "learning_rate": 1e-05,
-      "loss": 0.1572,
-      "step": 520
-    },
-    {
-      "epoch": 2.5,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.09873003512620926,
-      "eval_runtime": 2.8,
-      "eval_samples_per_second": 33.215,
-      "eval_steps_per_second": 8.572,
-      "step": 520
-    },
-    {
-      "epoch": 2.5961538461538463,
-      "grad_norm": 0.005777462385594845,
-      "learning_rate": 9.615384615384616e-06,
-      "loss": 0.0359,
-      "step": 540
-    },
-    {
-      "epoch": 2.5961538461538463,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.12076444178819656,
-      "eval_runtime": 2.8713,
-      "eval_samples_per_second": 32.389,
-      "eval_steps_per_second": 8.359,
-      "step": 540
-    },
-    {
-      "epoch": 2.6923076923076925,
-      "grad_norm": 0.053336434066295624,
-      "learning_rate": 9.230769230769232e-06,
-      "loss": 0.0737,
-      "step": 560
-    },
-    {
-      "epoch": 2.6923076923076925,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.09739954769611359,
-      "eval_runtime": 2.8488,
-      "eval_samples_per_second": 32.645,
-      "eval_steps_per_second": 8.425,
-      "step": 560
-    },
-    {
-      "epoch": 2.7884615384615383,
-      "grad_norm": 0.20622719824314117,
-      "learning_rate": 8.846153846153847e-06,
-      "loss": 0.1555,
-      "step": 580
-    },
-    {
-      "epoch": 2.7884615384615383,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.1363420933485031,
-      "eval_runtime": 2.7797,
-      "eval_samples_per_second": 33.456,
-      "eval_steps_per_second": 8.634,
-      "step": 580
-    },
-    {
-      "epoch": 2.8846153846153846,
-      "grad_norm": 0.006419941317290068,
-      "learning_rate": 8.461538461538462e-06,
-      "loss": 0.0928,
-      "step": 600
-    },
-    {
-      "epoch": 2.8846153846153846,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.06813652068376541,
-      "eval_runtime": 2.731,
-      "eval_samples_per_second": 34.053,
-      "eval_steps_per_second": 8.788,
-      "step": 600
-    },
-    {
-      "epoch": 2.980769230769231,
-      "grad_norm": 0.01808270625770092,
-      "learning_rate": 8.076923076923077e-06,
-      "loss": 0.0008,
-      "step": 620
-    },
-    {
-      "epoch": 2.980769230769231,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.06113424152135849,
-      "eval_runtime": 2.7355,
-      "eval_samples_per_second": 33.997,
-      "eval_steps_per_second": 8.773,
-      "step": 620
-    },
-    {
-      "epoch": 3.076923076923077,
-      "grad_norm": 0.01133087370544672,
-      "learning_rate": 7.692307692307694e-06,
-      "loss": 0.0606,
-      "step": 640
-    },
-    {
-      "epoch": 3.076923076923077,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.09787502139806747,
-      "eval_runtime": 2.6887,
-      "eval_samples_per_second": 34.589,
-      "eval_steps_per_second": 8.926,
-      "step": 640
-    },
-    {
-      "epoch": 3.173076923076923,
-      "grad_norm": 0.009248894639313221,
-      "learning_rate": 7.307692307692308e-06,
-      "loss": 0.0693,
-      "step": 660
-    },
-    {
-      "epoch": 3.173076923076923,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.07677934318780899,
-      "eval_runtime": 2.7198,
-      "eval_samples_per_second": 34.194,
-      "eval_steps_per_second": 8.824,
-      "step": 660
-    },
-    {
-      "epoch": 3.269230769230769,
-      "grad_norm": 0.00641106441617012,
-      "learning_rate": 6.923076923076923e-06,
-      "loss": 0.0005,
-      "step": 680
-    },
-    {
-      "epoch": 3.269230769230769,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.09253555536270142,
-      "eval_runtime": 2.7401,
-      "eval_samples_per_second": 33.94,
-      "eval_steps_per_second": 8.759,
-      "step": 680
-    },
-    {
-      "epoch": 3.3653846153846154,
-      "grad_norm": 0.015779003500938416,
-      "learning_rate": 6.538461538461539e-06,
-      "loss": 0.0006,
-      "step": 700
-    },
-    {
-      "epoch": 3.3653846153846154,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.09215600788593292,
-      "eval_runtime": 2.7734,
-      "eval_samples_per_second": 33.533,
-      "eval_steps_per_second": 8.654,
-      "step": 700
-    },
-    {
-      "epoch": 3.4615384615384617,
-      "grad_norm": 0.008339837193489075,
-      "learning_rate": 6.153846153846155e-06,
-      "loss": 0.0005,
-      "step": 720
-    },
-    {
-      "epoch": 3.4615384615384617,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.09074998646974564,
-      "eval_runtime": 2.7823,
-      "eval_samples_per_second": 33.426,
-      "eval_steps_per_second": 8.626,
-      "step": 720
-    },
-    {
-      "epoch": 3.5576923076923075,
-      "grad_norm": 0.01772761344909668,
-      "learning_rate": 5.769230769230769e-06,
-      "loss": 0.0004,
-      "step": 740
-    },
-    {
-      "epoch": 3.5576923076923075,
-      "eval_accuracy": 0.967741935483871,
-      "eval_loss": 0.09229286760091782,
-      "eval_runtime": 2.7698,
-      "eval_samples_per_second": 33.577,
-      "eval_steps_per_second": 8.665,
-      "step": 740
-    },
-    {
-      "epoch": 3.6538461538461537,
-      "grad_norm": 25.846527099609375,
-      "learning_rate": 5.384615384615385e-06,
-      "loss": 0.056,
-      "step": 760
-    },
-    {
-      "epoch": 3.6538461538461537,
-      "eval_accuracy": 0.956989247311828,
-      "eval_loss": 0.09057936072349548,
-      "eval_runtime": 2.7524,
-      "eval_samples_per_second": 33.789,
-      "eval_steps_per_second": 8.72,
-      "step": 760
-    },
-    {
-      "epoch": 3.75,
-      "grad_norm": 0.007363874930888414,
-      "learning_rate": 5e-06,
-      "loss": 0.0006,
-      "step": 780
-    },
-    {
-      "epoch": 3.75,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.09129434823989868,
-      "eval_runtime": 2.7344,
-      "eval_samples_per_second": 34.011,
-      "eval_steps_per_second": 8.777,
-      "step": 780
-    },
-    {
-      "epoch": 3.8461538461538463,
-      "grad_norm": 0.0785047858953476,
-      "learning_rate": 4.615384615384616e-06,
-      "loss": 0.056,
-      "step": 800
-    },
-    {
-      "epoch": 3.8461538461538463,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.11733542382717133,
-      "eval_runtime": 2.7493,
-      "eval_samples_per_second": 33.826,
-      "eval_steps_per_second": 8.729,
-      "step": 800
-    },
-    {
-      "epoch": 3.9423076923076925,
-      "grad_norm": 0.6651670336723328,
-      "learning_rate": 4.230769230769231e-06,
-      "loss": 0.0005,
-      "step": 820
-    },
-    {
-      "epoch": 3.9423076923076925,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.1340881735086441,
-      "eval_runtime": 2.7478,
-      "eval_samples_per_second": 33.846,
-      "eval_steps_per_second": 8.734,
-      "step": 820
-    },
-    {
-      "epoch": 4.038461538461538,
-      "grad_norm": 0.005213054362684488,
-      "learning_rate": 3.846153846153847e-06,
-      "loss": 0.0004,
-      "step": 840
-    },
-    {
-      "epoch": 4.038461538461538,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.11991964280605316,
-      "eval_runtime": 2.7527,
-      "eval_samples_per_second": 33.784,
-      "eval_steps_per_second": 8.719,
-      "step": 840
-    },
-    {
-      "epoch": 4.134615384615385,
-      "grad_norm": 0.005048852413892746,
-      "learning_rate": 3.4615384615384617e-06,
-      "loss": 0.0004,
-      "step": 860
-    },
-    {
-      "epoch": 4.134615384615385,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.11612516641616821,
-      "eval_runtime": 2.757,
-      "eval_samples_per_second": 33.732,
-      "eval_steps_per_second": 8.705,
-      "step": 860
-    },
-    {
-      "epoch": 4.230769230769231,
-      "grad_norm": 0.0033646891824901104,
-      "learning_rate": 3.0769230769230774e-06,
-      "loss": 0.0003,
-      "step": 880
-    },
-    {
-      "epoch": 4.230769230769231,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.11559664458036423,
-      "eval_runtime": 2.7327,
-      "eval_samples_per_second": 34.032,
-      "eval_steps_per_second": 8.783,
-      "step": 880
-    },
-    {
-      "epoch": 4.326923076923077,
-      "grad_norm": 0.004703676328063011,
-      "learning_rate": 2.6923076923076923e-06,
-      "loss": 0.0385,
-      "step": 900
-    },
-    {
-      "epoch": 4.326923076923077,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.08591005206108093,
-      "eval_runtime": 2.714,
-      "eval_samples_per_second": 34.266,
-      "eval_steps_per_second": 8.843,
-      "step": 900
-    },
-    {
-      "epoch": 4.423076923076923,
-      "grad_norm": 0.011793052777647972,
-      "learning_rate": 2.307692307692308e-06,
-      "loss": 0.0028,
-      "step": 920
-    },
-    {
-      "epoch": 4.423076923076923,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.09606783837080002,
-      "eval_runtime": 2.7185,
-      "eval_samples_per_second": 34.21,
-      "eval_steps_per_second": 8.828,
-      "step": 920
-    },
-    {
-      "epoch": 4.519230769230769,
-      "grad_norm": 0.009880756959319115,
-      "learning_rate": 1.9230769230769234e-06,
-      "loss": 0.0003,
-      "step": 940
-    },
-    {
-      "epoch": 4.519230769230769,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.1021433100104332,
-      "eval_runtime": 2.7534,
-      "eval_samples_per_second": 33.777,
-      "eval_steps_per_second": 8.717,
-      "step": 940
-    },
-    {
-      "epoch": 4.615384615384615,
-      "grad_norm": 0.0028292695060372353,
-      "learning_rate": 1.5384615384615387e-06,
-      "loss": 0.0014,
-      "step": 960
-    },
-    {
-      "epoch": 4.615384615384615,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.10712174326181412,
-      "eval_runtime": 2.7663,
-      "eval_samples_per_second": 33.619,
-      "eval_steps_per_second": 8.676,
-      "step": 960
-    },
-    {
-      "epoch": 4.711538461538462,
-      "grad_norm": 0.003742179134860635,
-      "learning_rate": 1.153846153846154e-06,
-      "loss": 0.0003,
-      "step": 980
-    },
-    {
-      "epoch": 4.711538461538462,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.10891132056713104,
-      "eval_runtime": 2.7626,
-      "eval_samples_per_second": 33.664,
-      "eval_steps_per_second": 8.688,
-      "step": 980
-    },
-    {
-      "epoch": 4.8076923076923075,
-      "grad_norm": 0.006350115407258272,
-      "learning_rate": 7.692307692307694e-07,
-      "loss": 0.0003,
-      "step": 1000
-    },
-    {
-      "epoch": 4.8076923076923075,
-      "eval_accuracy": 0.978494623655914,
-      "eval_loss": 0.10818465054035187,
-      "eval_runtime": 2.7504,
-      "eval_samples_per_second": 33.813,
-      "eval_steps_per_second": 8.726,
-      "step": 1000
     }
   ],
   "logging_steps": 20,
@@ -826,7 +426,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 971631613563540.0,
+  "total_flos": 480317453780160.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:27fa1e1529bd8193e84d4f9fcaca6c5833806923edc0d3e169c3b554b74ed8ed
+oid sha256:c4d9f02b57c34a8613828c536e382079e3121c1169bdb4aae43dee6df2c3099d
 size 5112