Training in progress, step 1000, checkpoint

checkpoint-1000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:def9ace5dae2dfaa12529cbdfcadf42b3c354925357cd060eab98b57da0a9e43
+oid sha256:17905a80aff01971287c50b4975452d184a531d66058fcbdf43415cf591a0fca
 size 932401285

checkpoint-1000/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e740544dbc8f4c131bbb9dbe3b1d731728fdcce66fb832a589cf58273105913
+oid sha256:41843b81f57d0f200a674feb879b70c8817af240929ee7617243b793b2b2e17f
 size 466205145

checkpoint-1000/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:11d0b2674492d14792eb3c9d600666ecb25f035d03b0399d0ac97f8400d34237
+oid sha256:9bd5412e9f240ee6cc27d51943a8f912bd03270f4fb7b31a33cfca544316ee5b
 size 14575

checkpoint-1000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 1.1078879833221436,
+  "best_metric": 1.1077626943588257,
   "best_model_checkpoint": "/kaggle/output/checkpoint-1000",
-  "epoch": 16.0,
+  "epoch": 0.04074315514993481,
   "eval_steps": 1000,
   "global_step": 1000,
   "is_hyper_param_search": false,
@@ -9,30 +9,30 @@
   "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
     }
   ],
   "logging_steps": 1000,
   "max_steps": 10000000,
-  "num_train_epochs": 161291,
+  "num_train_epochs": 408,
   "save_steps": 1000,
   "total_flos": 2090355130368000.0,
   "trial_name": null,

checkpoint-1000/training_args.bin

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