Training in progress, step 2000, checkpoint

checkpoint-2000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd40d9fc39d514e206a41dd6f56b3f3008ee39fff7e003ec639420e698e57ad5
+oid sha256:d22b85a7263581b2f995229898bfa3fbff963525eb21b6ff27842bf870f14d11
 size 932401285

checkpoint-2000/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:793a4df80b66224da08e68b7a83fff445266efb565820fc2d0c0e0f23738e4ff
+oid sha256:45866a9412bf8db18f4fb57252b2c9109cf8bbea03bebf71de24a6bbc10c5a00
 size 466205145

checkpoint-2000/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:125f7f290df775248359bf13a2e7ca2ed02ca18d1d0ee142aa6142af8f9cc55d
+oid sha256:1df9fb5d9a26bcf3d16f43a9e2e707cdf1b1348f36924b93d279819e285489ed
 size 14575

checkpoint-2000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 1.1078879833221436,
+  "best_metric": 1.1077626943588257,
   "best_model_checkpoint": "/kaggle/output/checkpoint-1000",
-  "epoch": 32.0,
+  "epoch": 0.08148631029986962,
   "eval_steps": 1000,
   "global_step": 2000,
   "is_hyper_param_search": false,
@@ -9,45 +9,45 @@
   "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
     }
   ],
   "logging_steps": 1000,
   "max_steps": 10000000,
-  "num_train_epochs": 161291,
+  "num_train_epochs": 408,
   "save_steps": 1000,
   "total_flos": 4180710260736000.0,
   "trial_name": null,

checkpoint-2000/training_args.bin

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