Training in progress, step 15000, checkpoint

checkpoint-15000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:74b48bbb6d39f9f404cad994fe19d43959944de85a6d908672cd1196daab51f8
+oid sha256:63a12a7316c5b07e5f3ecb4ae9282e13dbadae9437c95a6d1b9ac55c3af5169b
 size 932401285

checkpoint-15000/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:51b76e40d6f337992da1d42594b8f83c9172b84c8ea36f59911781422ec7dd9b
+oid sha256:ef50775206cd307159416ec20146a07b5d81fc556295d4f2cb4ebbf28f12c28c
 size 466205145

checkpoint-15000/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fea505d3bb2cc8e103041c009ac11cec0d896dbfe4c76b96d22f081a5d99d8b8
+oid sha256:f1a713585e10ca5328b118b85cc0cd1d602b0bee2bece307a8b1b8aac63ef26b
 size 14575

checkpoint-15000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:226555fde61939e5ec94e16de92c4a9b1ef7a27af8c9ddcdc8e335cc55b7bfcc
+oid sha256:a65fb4364749449147c94e9d64805571fea126ee59e61672a85867dcac3af9c8
 size 627

checkpoint-15000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 1.1056212186813354,
-  "best_model_checkpoint": "/kaggle/output/checkpoint-6000",
-  "epoch": 240.0,
+  "best_metric": 1.0886671543121338,
+  "best_model_checkpoint": "/kaggle/output/checkpoint-15000",
+  "epoch": 0.6111473272490222,
   "eval_steps": 1000,
   "global_step": 15000,
   "is_hyper_param_search": false,
@@ -9,240 +9,240 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.02,
+      "epoch": 0.0,
       "learning_rate": 2.7777777777777777e-11,
-      "loss": 1.5477,
+      "loss": 1.1383,
       "step": 1
     },
     {
-      "epoch": 16.0,
+      "epoch": 0.04,
       "learning_rate": 2.7750000000000004e-08,
-      "loss": 1.2028,
+      "loss": 1.1424,
       "step": 1000
     },
     {
-      "epoch": 16.0,
-      "eval_accuracy": 0.3387225548902196,
-      "eval_loss": 1.1078879833221436,
-      "eval_runtime": 54.1057,
-      "eval_samples_per_second": 92.596,
-      "eval_steps_per_second": 11.588,
+      "epoch": 0.04,
+      "eval_accuracy": 0.32375249500998005,
+      "eval_loss": 1.1077626943588257,
+      "eval_runtime": 54.8633,
+      "eval_samples_per_second": 91.318,
+      "eval_steps_per_second": 11.428,
       "step": 1000
     },
     {
-      "epoch": 32.0,
+      "epoch": 0.08,
       "learning_rate": 5.5527777777777784e-08,
-      "loss": 1.1152,
+      "loss": 1.1244,
       "step": 2000
     },
     {
-      "epoch": 32.0,
-      "eval_accuracy": 0.3349301397205589,
-      "eval_loss": 1.1080944538116455,
-      "eval_runtime": 54.2029,
-      "eval_samples_per_second": 92.43,
-      "eval_steps_per_second": 11.568,
+      "epoch": 0.08,
+      "eval_accuracy": 0.33652694610778444,
+      "eval_loss": 1.1080161333084106,
+      "eval_runtime": 54.7384,
+      "eval_samples_per_second": 91.526,
+      "eval_steps_per_second": 11.454,
       "step": 2000
     },
     {
-      "epoch": 48.0,
+      "epoch": 0.12,
       "learning_rate": 8.327777777777778e-08,
-      "loss": 1.1036,
+      "loss": 1.1228,
       "step": 3000
     },
     {
-      "epoch": 48.0,
-      "eval_accuracy": 0.3377245508982036,
-      "eval_loss": 1.1081210374832153,
-      "eval_runtime": 54.0467,
-      "eval_samples_per_second": 92.698,
-      "eval_steps_per_second": 11.601,
+      "epoch": 0.12,
+      "eval_accuracy": 0.34331337325349304,
+      "eval_loss": 1.1084064245224,
+      "eval_runtime": 54.7948,
+      "eval_samples_per_second": 91.432,
+      "eval_steps_per_second": 11.443,
       "step": 3000
     },
     {
-      "epoch": 64.0,
+      "epoch": 0.16,
       "learning_rate": 1.1105555555555557e-07,
-      "loss": 1.0895,
+      "loss": 1.1216,
       "step": 4000
     },
     {
-      "epoch": 64.0,
-      "eval_accuracy": 0.33912175648702597,
-      "eval_loss": 1.1072543859481812,
-      "eval_runtime": 54.1318,
-      "eval_samples_per_second": 92.552,
-      "eval_steps_per_second": 11.583,
+      "epoch": 0.16,
+      "eval_accuracy": 0.3385229540918164,
+      "eval_loss": 1.1014840602874756,
+      "eval_runtime": 54.8508,
+      "eval_samples_per_second": 91.339,
+      "eval_steps_per_second": 11.431,
       "step": 4000
     },
     {
-      "epoch": 80.0,
+      "epoch": 0.2,
       "learning_rate": 1.3880555555555558e-07,
-      "loss": 1.0748,
+      "loss": 1.1181,
       "step": 5000
     },
     {
-      "epoch": 80.0,
-      "eval_accuracy": 0.3435129740518962,
-      "eval_loss": 1.1071943044662476,
-      "eval_runtime": 54.4121,
-      "eval_samples_per_second": 92.075,
-      "eval_steps_per_second": 11.523,
+      "epoch": 0.2,
+      "eval_accuracy": 0.33073852295409184,
+      "eval_loss": 1.1008135080337524,
+      "eval_runtime": 54.8304,
+      "eval_samples_per_second": 91.373,
+      "eval_steps_per_second": 11.435,
       "step": 5000
     },
     {
-      "epoch": 96.0,
+      "epoch": 0.24,
       "learning_rate": 1.6658333333333335e-07,
-      "loss": 1.055,
+      "loss": 1.1132,
       "step": 6000
     },
     {
-      "epoch": 96.0,
-      "eval_accuracy": 0.3459081836327345,
-      "eval_loss": 1.1056212186813354,
-      "eval_runtime": 54.092,
-      "eval_samples_per_second": 92.62,
-      "eval_steps_per_second": 11.591,
+      "epoch": 0.24,
+      "eval_accuracy": 0.3520958083832335,
+      "eval_loss": 1.0993762016296387,
+      "eval_runtime": 54.8804,
+      "eval_samples_per_second": 91.289,
+      "eval_steps_per_second": 11.425,
       "step": 6000
     },
     {
-      "epoch": 112.0,
+      "epoch": 0.29,
       "learning_rate": 1.9433333333333334e-07,
-      "loss": 1.0352,
+      "loss": 1.1113,
       "step": 7000
     },
     {
-      "epoch": 112.0,
-      "eval_accuracy": 0.3469061876247505,
-      "eval_loss": 1.1121071577072144,
-      "eval_runtime": 53.9318,
-      "eval_samples_per_second": 92.895,
-      "eval_steps_per_second": 11.626,
+      "epoch": 0.29,
+      "eval_accuracy": 0.3530938123752495,
+      "eval_loss": 1.0965770483016968,
+      "eval_runtime": 54.8881,
+      "eval_samples_per_second": 91.277,
+      "eval_steps_per_second": 11.423,
       "step": 7000
     },
     {
-      "epoch": 128.0,
+      "epoch": 0.33,
       "learning_rate": 2.2211111111111114e-07,
-      "loss": 1.0085,
+      "loss": 1.1111,
       "step": 8000
     },
     {
-      "epoch": 128.0,
-      "eval_accuracy": 0.3534930139720559,
-      "eval_loss": 1.1201248168945312,
-      "eval_runtime": 54.0094,
-      "eval_samples_per_second": 92.762,
-      "eval_steps_per_second": 11.609,
+      "epoch": 0.33,
+      "eval_accuracy": 0.35708582834331337,
+      "eval_loss": 1.094658613204956,
+      "eval_runtime": 54.8233,
+      "eval_samples_per_second": 91.384,
+      "eval_steps_per_second": 11.437,
       "step": 8000
     },
     {
-      "epoch": 144.0,
+      "epoch": 0.37,
       "learning_rate": 2.4986111111111113e-07,
-      "loss": 0.9772,
+      "loss": 1.109,
       "step": 9000
     },
     {
-      "epoch": 144.0,
-      "eval_accuracy": 0.3568862275449102,
-      "eval_loss": 1.1434961557388306,
-      "eval_runtime": 54.0639,
-      "eval_samples_per_second": 92.668,
-      "eval_steps_per_second": 11.597,
+      "epoch": 0.37,
+      "eval_accuracy": 0.34191616766467064,
+      "eval_loss": 1.106990933418274,
+      "eval_runtime": 54.9095,
+      "eval_samples_per_second": 91.241,
+      "eval_steps_per_second": 11.419,
       "step": 9000
     },
     {
-      "epoch": 160.0,
-      "learning_rate": 2.7761111111111115e-07,
-      "loss": 0.9323,
+      "epoch": 0.41,
+      "learning_rate": 2.776388888888889e-07,
+      "loss": 1.1036,
       "step": 10000
     },
     {
-      "epoch": 160.0,
-      "eval_accuracy": 0.36087824351297404,
-      "eval_loss": 1.1742260456085205,
-      "eval_runtime": 54.7624,
-      "eval_samples_per_second": 91.486,
-      "eval_steps_per_second": 11.449,
+      "epoch": 0.41,
+      "eval_accuracy": 0.37584830339321357,
+      "eval_loss": 1.0930211544036865,
+      "eval_runtime": 54.9067,
+      "eval_samples_per_second": 91.246,
+      "eval_steps_per_second": 11.419,
       "step": 10000
     },
     {
-      "epoch": 176.0,
-      "learning_rate": 3.053888888888889e-07,
-      "loss": 0.8779,
+      "epoch": 0.45,
+      "learning_rate": 3.0541666666666667e-07,
+      "loss": 1.1045,
       "step": 11000
     },
     {
-      "epoch": 176.0,
-      "eval_accuracy": 0.3568862275449102,
-      "eval_loss": 1.2272026538848877,
-      "eval_runtime": 54.0934,
-      "eval_samples_per_second": 92.618,
-      "eval_steps_per_second": 11.591,
+      "epoch": 0.45,
+      "eval_accuracy": 0.3652694610778443,
+      "eval_loss": 1.092846393585205,
+      "eval_runtime": 54.8964,
+      "eval_samples_per_second": 91.263,
+      "eval_steps_per_second": 11.422,
       "step": 11000
     },
     {
-      "epoch": 192.0,
-      "learning_rate": 3.331388888888889e-07,
-      "loss": 0.8276,
+      "epoch": 0.49,
+      "learning_rate": 3.3319444444444444e-07,
+      "loss": 1.1024,
       "step": 12000
     },
     {
-      "epoch": 192.0,
-      "eval_accuracy": 0.3562874251497006,
-      "eval_loss": 1.3099753856658936,
-      "eval_runtime": 54.3996,
-      "eval_samples_per_second": 92.096,
-      "eval_steps_per_second": 11.526,
+      "epoch": 0.49,
+      "eval_accuracy": 0.39261477045908183,
+      "eval_loss": 1.089038372039795,
+      "eval_runtime": 54.9763,
+      "eval_samples_per_second": 91.13,
+      "eval_steps_per_second": 11.405,
       "step": 12000
     },
     {
-      "epoch": 208.0,
-      "learning_rate": 3.608888888888889e-07,
-      "loss": 0.7716,
+      "epoch": 0.53,
+      "learning_rate": 3.6094444444444446e-07,
+      "loss": 1.1007,
       "step": 13000
     },
     {
-      "epoch": 208.0,
-      "eval_accuracy": 0.3588822355289421,
-      "eval_loss": 1.4230743646621704,
-      "eval_runtime": 54.193,
-      "eval_samples_per_second": 92.447,
-      "eval_steps_per_second": 11.57,
+      "epoch": 0.53,
+      "eval_accuracy": 0.34311377245508984,
+      "eval_loss": 1.0933948755264282,
+      "eval_runtime": 54.9285,
+      "eval_samples_per_second": 91.209,
+      "eval_steps_per_second": 11.415,
       "step": 13000
     },
     {
-      "epoch": 224.0,
-      "learning_rate": 3.886666666666667e-07,
-      "loss": 0.7162,
+      "epoch": 0.57,
+      "learning_rate": 3.8872222222222223e-07,
+      "loss": 1.0985,
       "step": 14000
     },
     {
-      "epoch": 224.0,
-      "eval_accuracy": 0.35568862275449104,
-      "eval_loss": 1.547025442123413,
-      "eval_runtime": 54.8911,
-      "eval_samples_per_second": 91.272,
-      "eval_steps_per_second": 11.423,
+      "epoch": 0.57,
+      "eval_accuracy": 0.36367265469061877,
+      "eval_loss": 1.09434974193573,
+      "eval_runtime": 54.8032,
+      "eval_samples_per_second": 91.418,
+      "eval_steps_per_second": 11.441,
       "step": 14000
     },
     {
-      "epoch": 240.0,
-      "learning_rate": 4.164166666666667e-07,
-      "loss": 0.6447,
+      "epoch": 0.61,
+      "learning_rate": 4.1650000000000006e-07,
+      "loss": 1.0988,
       "step": 15000
     },
     {
-      "epoch": 240.0,
-      "eval_accuracy": 0.35528942115768464,
-      "eval_loss": 1.6746374368667603,
-      "eval_runtime": 55.1232,
-      "eval_samples_per_second": 90.887,
-      "eval_steps_per_second": 11.375,
+      "epoch": 0.61,
+      "eval_accuracy": 0.39481037924151696,
+      "eval_loss": 1.0886671543121338,
+      "eval_runtime": 54.9221,
+      "eval_samples_per_second": 91.22,
+      "eval_steps_per_second": 11.416,
       "step": 15000
     }
   ],
   "logging_steps": 1000,
   "max_steps": 10000000,
-  "num_train_epochs": 161291,
+  "num_train_epochs": 408,
   "save_steps": 1000,
   "total_flos": 3.135532695552e+16,
   "trial_name": null,

checkpoint-15000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6c09d25f832c6c38e287ec7a7fe1b191faf603815bb3b91e32f16768a2d7d3ff
+oid sha256:a9a52921b15d7eeb53b158a9c0f943d88288ccd9232004774ffdf3a600a048d5
 size 4091