Training in progress, step 4000, checkpoint

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:927cc3a6954f0ffa37523365e836fc8d0e054ba9e9b62e4efeef208689ca1e69
+oid sha256:e885cac5c061624d6bc12dc0e231dbf23d716a260ebf59bb508281da295fe021
 size 499730128

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:27f5349faf73282fae15b6197ce207ae54649a387b8a2ac9d9d6289a07744cb1
-size 1442699
+oid sha256:e1d0700dd49d8512365f4fe50959ea0dbc3be2ffc5b05812fda8483189b4e5a4
+size 812034507

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6d920d35377b6218b0a182ca529dc7f210a3398080e4578011b7878234683794
+oid sha256:9e06dfbf969388fb27efc692631d298abe0d1c57738a0c64598fe664b1ac46b1
 size 14645

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b716fb1dcbc77ddc6a0ae7c4c98d559e36f1986fff006ef6982bd75247833235
+oid sha256:76cf98a8cde877f1e49a9e4b884c4dce3d51bb4035429d464e74acc35ddca024
 size 1465

last-checkpoint/trainer_state.json

@@ -2,170 +2,328 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.04,
+  "epoch": 0.08,
   "eval_steps": 1000,
-  "global_step": 2000,
+  "global_step": 4000,
   "is_hyper_param_search": false,
-  "is_local_process_zero": false,
+  "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.002,
-      "grad_norm": 19.660429000854492,
+      "grad_norm": 1.2400550842285156,
       "learning_rate": 7.920000000000001e-05,
-      "loss": 67.8543,
+      "loss": 8.362694702148438,
       "step": 100
     },
     {
       "epoch": 0.004,
-      "grad_norm": 24.802082061767578,
+      "grad_norm": 2.1569149494171143,
       "learning_rate": 0.00015920000000000002,
-      "loss": 54.1444,
+      "loss": 6.592593383789063,
       "step": 200
     },
     {
       "epoch": 0.006,
-      "grad_norm": 16.059066772460938,
+      "grad_norm": 4.070979595184326,
       "learning_rate": 0.00023920000000000001,
-      "loss": 51.5921,
+      "loss": 6.208985595703125,
       "step": 300
     },
     {
       "epoch": 0.008,
-      "grad_norm": 27.705520629882812,
+      "grad_norm": 4.392622470855713,
       "learning_rate": 0.0003192,
-      "loss": 50.0536,
+      "loss": 5.921519775390625,
       "step": 400
     },
     {
       "epoch": 0.01,
-      "grad_norm": 25.79052734375,
+      "grad_norm": 4.653286457061768,
       "learning_rate": 0.0003992,
-      "loss": 48.6987,
+      "loss": 5.658856201171875,
       "step": 500
     },
     {
       "epoch": 0.012,
-      "grad_norm": 14.82126522064209,
+      "grad_norm": 2.6610119342803955,
       "learning_rate": 0.00047920000000000005,
-      "loss": 47.3182,
+      "loss": 5.450134887695312,
       "step": 600
     },
     {
       "epoch": 0.014,
-      "grad_norm": 12.00988483428955,
+      "grad_norm": 2.6298203468322754,
       "learning_rate": 0.0005592,
-      "loss": 46.6574,
+      "loss": 5.260469360351562,
       "step": 700
     },
     {
       "epoch": 0.016,
-      "grad_norm": 32.48664474487305,
+      "grad_norm": 1.681970238685608,
       "learning_rate": 0.0006392,
-      "loss": 46.0747,
+      "loss": 5.112535095214843,
       "step": 800
     },
     {
       "epoch": 0.018,
-      "grad_norm": 25.646282196044922,
+      "grad_norm": 1.5673184394836426,
       "learning_rate": 0.0007191999999999999,
-      "loss": 45.6098,
+      "loss": 5.0076907348632815,
       "step": 900
     },
     {
       "epoch": 0.02,
-      "grad_norm": 16.071203231811523,
+      "grad_norm": 1.6654455661773682,
       "learning_rate": 0.0007992,
-      "loss": 44.8537,
+      "loss": 4.9159228515625,
       "step": 1000
     },
     {
       "epoch": 0.02,
-      "eval_accuracy": 0.18606457925636008,
-      "eval_loss": 44.18140411376953,
-      "eval_runtime": 14.581,
-      "eval_samples_per_second": 8.573,
-      "eval_steps_per_second": 0.137,
+      "eval_accuracy": 0.2274637964774951,
+      "eval_loss": 4.8702616691589355,
+      "eval_runtime": 5.1375,
+      "eval_samples_per_second": 194.646,
+      "eval_steps_per_second": 1.557,
       "step": 1000
     },
     {
       "epoch": 0.022,
-      "grad_norm": 14.277446746826172,
+      "grad_norm": 1.3562606573104858,
       "learning_rate": 0.0008792,
-      "loss": 44.223,
+      "loss": 4.811459045410157,
       "step": 1100
     },
     {
       "epoch": 0.024,
-      "grad_norm": 9.809870719909668,
+      "grad_norm": 1.1230770349502563,
       "learning_rate": 0.0009592000000000001,
-      "loss": 43.5719,
+      "loss": 4.759424438476563,
       "step": 1200
     },
     {
       "epoch": 0.026,
-      "grad_norm": 16.994014739990234,
+      "grad_norm": 0.9026045203208923,
       "learning_rate": 0.0010391999999999999,
-      "loss": 43.0323,
+      "loss": 4.691336669921875,
       "step": 1300
     },
     {
       "epoch": 0.028,
-      "grad_norm": 8.082005500793457,
+      "grad_norm": 0.8914104104042053,
       "learning_rate": 0.0011192,
-      "loss": 42.3631,
+      "loss": 4.6415283203125,
       "step": 1400
     },
     {
       "epoch": 0.03,
-      "grad_norm": 7.080636024475098,
+      "grad_norm": 0.8752658367156982,
       "learning_rate": 0.0011992,
-      "loss": 42.0285,
+      "loss": 4.560531311035156,
       "step": 1500
     },
     {
       "epoch": 0.032,
-      "grad_norm": 8.547815322875977,
+      "grad_norm": 0.7435582280158997,
       "learning_rate": 0.0012791999999999999,
-      "loss": 42.6829,
+      "loss": 4.516304016113281,
       "step": 1600
     },
     {
       "epoch": 0.034,
-      "grad_norm": 7.541939735412598,
+      "grad_norm": 0.7041499018669128,
       "learning_rate": 0.0013592,
-      "loss": 41.9796,
+      "loss": 4.48115234375,
       "step": 1700
     },
     {
       "epoch": 0.036,
-      "grad_norm": 6.902923583984375,
+      "grad_norm": 0.5327283143997192,
       "learning_rate": 0.0014392,
-      "loss": 41.5016,
+      "loss": 4.429443969726562,
       "step": 1800
     },
     {
       "epoch": 0.038,
-      "grad_norm": 7.68479061126709,
+      "grad_norm": 0.45207348465919495,
       "learning_rate": 0.0015192,
-      "loss": 41.2607,
+      "loss": 4.37362060546875,
       "step": 1900
     },
     {
       "epoch": 0.04,
-      "grad_norm": 8.936702728271484,
+      "grad_norm": 0.5152547359466553,
       "learning_rate": 0.0015992,
-      "loss": 40.7545,
+      "loss": 4.377236938476562,
       "step": 2000
     },
     {
       "epoch": 0.04,
-      "eval_accuracy": 0.21657142857142858,
-      "eval_loss": 40.417686462402344,
-      "eval_runtime": 3.051,
-      "eval_samples_per_second": 40.97,
-      "eval_steps_per_second": 0.656,
+      "eval_accuracy": 0.2694227005870842,
+      "eval_loss": 4.3582682609558105,
+      "eval_runtime": 5.3878,
+      "eval_samples_per_second": 185.606,
+      "eval_steps_per_second": 1.485,
       "step": 2000
+    },
+    {
+      "epoch": 0.042,
+      "grad_norm": 0.45765218138694763,
+      "learning_rate": 0.0016792,
+      "loss": 4.341053161621094,
+      "step": 2100
+    },
+    {
+      "epoch": 0.044,
+      "grad_norm": 0.3776390254497528,
+      "learning_rate": 0.0017592,
+      "loss": 4.304255676269531,
+      "step": 2200
+    },
+    {
+      "epoch": 0.046,
+      "grad_norm": 0.48194995522499084,
+      "learning_rate": 0.0018392,
+      "loss": 4.308780822753906,
+      "step": 2300
+    },
+    {
+      "epoch": 0.048,
+      "grad_norm": 0.40603286027908325,
+      "learning_rate": 0.0019192,
+      "loss": 4.272806396484375,
+      "step": 2400
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.41386568546295166,
+      "learning_rate": 0.0019992,
+      "loss": 4.267073974609375,
+      "step": 2500
+    },
+    {
+      "epoch": 0.052,
+      "grad_norm": 0.4916844666004181,
+      "learning_rate": 0.001999978563623903,
+      "loss": 4.220490417480469,
+      "step": 2600
+    },
+    {
+      "epoch": 0.054,
+      "grad_norm": 0.37477609515190125,
+      "learning_rate": 0.0019999133871331223,
+      "loss": 4.224794006347656,
+      "step": 2700
+    },
+    {
+      "epoch": 0.056,
+      "grad_norm": 0.405086874961853,
+      "learning_rate": 0.0019998044711915177,
+      "loss": 4.185610961914063,
+      "step": 2800
+    },
+    {
+      "epoch": 0.058,
+      "grad_norm": 0.334581196308136,
+      "learning_rate": 0.0019996518205634257,
+      "loss": 4.165193176269531,
+      "step": 2900
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.3760662376880646,
+      "learning_rate": 0.0019994554419262797,
+      "loss": 4.161107177734375,
+      "step": 3000
+    },
+    {
+      "epoch": 0.06,
+      "eval_accuracy": 0.28686888454011744,
+      "eval_loss": 4.15764856338501,
+      "eval_runtime": 5.1036,
+      "eval_samples_per_second": 195.938,
+      "eval_steps_per_second": 1.568,
+      "step": 3000
+    },
+    {
+      "epoch": 0.062,
+      "grad_norm": 0.38909581303596497,
+      "learning_rate": 0.001999215343870317,
+      "loss": 4.1489105224609375,
+      "step": 3100
+    },
+    {
+      "epoch": 0.064,
+      "grad_norm": 0.4108506143093109,
+      "learning_rate": 0.0019989315368982054,
+      "loss": 4.123897705078125,
+      "step": 3200
+    },
+    {
+      "epoch": 0.066,
+      "grad_norm": 0.27484413981437683,
+      "learning_rate": 0.0019986040334245797,
+      "loss": 4.101358032226562,
+      "step": 3300
+    },
+    {
+      "epoch": 0.068,
+      "grad_norm": 0.29958581924438477,
+      "learning_rate": 0.001998232847775504,
+      "loss": 4.10423583984375,
+      "step": 3400
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.29483968019485474,
+      "learning_rate": 0.0019978179961878404,
+      "loss": 4.078011169433593,
+      "step": 3500
+    },
+    {
+      "epoch": 0.072,
+      "grad_norm": 0.36317598819732666,
+      "learning_rate": 0.001997359496808541,
+      "loss": 4.076428527832031,
+      "step": 3600
+    },
+    {
+      "epoch": 0.074,
+      "grad_norm": 0.34998658299446106,
+      "learning_rate": 0.001996857369693855,
+      "loss": 4.052371826171875,
+      "step": 3700
+    },
+    {
+      "epoch": 0.076,
+      "grad_norm": 0.31070172786712646,
+      "learning_rate": 0.0019963116368084486,
+      "loss": 4.059148559570312,
+      "step": 3800
+    },
+    {
+      "epoch": 0.078,
+      "grad_norm": 0.2655385732650757,
+      "learning_rate": 0.001995722322024446,
+      "loss": 4.039622802734375,
+      "step": 3900
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.4650065302848816,
+      "learning_rate": 0.001995089451120385,
+      "loss": 4.011789855957031,
+      "step": 4000
+    },
+    {
+      "epoch": 0.08,
+      "eval_accuracy": 0.29829745596868884,
+      "eval_loss": 4.0341691970825195,
+      "eval_runtime": 5.1511,
+      "eval_samples_per_second": 194.133,
+      "eval_steps_per_second": 1.553,
+      "step": 4000
     }
   ],
   "logging_steps": 100,
@@ -186,7 +344,7 @@
     }
   },
   "total_flos": 0.0,
-  "train_batch_size": 8,
+  "train_batch_size": 128,
   "trial_name": null,
   "trial_params": null
 }

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:360066b4a5f1cc8c89941d389baaf9061071c4da2a624ff0fb9b3a9211c62ff7
-size 5905
+oid sha256:3ab34ea13b0e64af4c31a655731e444610b50d6a753b685ccf0d645cea5f76a3
+size 5265