Training in progress, step 1000, checkpoint

checkpoint-1000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:720e50b0a438112b33a0250a65da51d19fee5fff1fb1f218c39ca89b8aad8ffd
+oid sha256:61ba01fb9f6d664a23d1ab0a7974e18f8814a03665d747041ca0eb8c3c2ba749
 size 932401285

checkpoint-1000/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d1d1ac85674789381a4362a9c72fed5bdb9ed94a3db342f39e46d06a9fa86ccc
+oid sha256:f8885f59e181e06ccad17d474f05be9e5aaf229a32757d53fdcf7df4ca5f9c18
 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": 1.1557353734970093,
+  "best_metric": 1.1529844999313354,
   "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": 1.2395,
+      "loss": 1.029,
       "step": 1
     },
     {
-      "epoch": 16.0,
+      "epoch": 0.04,
       "learning_rate": 2.7638888888888893e-08,
-      "loss": 1.1422,
+      "loss": 1.181,
       "step": 1000
     },
     {
-      "epoch": 16.0,
-      "eval_accuracy": 0.33073852295409184,
-      "eval_loss": 1.1557353734970093,
-      "eval_runtime": 54.1805,
-      "eval_samples_per_second": 92.469,
-      "eval_steps_per_second": 11.572,
+      "epoch": 0.04,
+      "eval_accuracy": 0.3273453093812375,
+      "eval_loss": 1.1529844999313354,
+      "eval_runtime": 54.2837,
+      "eval_samples_per_second": 92.293,
+      "eval_steps_per_second": 11.55,
       "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:cb2cd1b7e648a386ccb79e47a451842e11001981c4363e993b678e0818db5cab
+oid sha256:5520446b18da09f4a43173b75839457981a344c0e709fd4bbcb249e3f60c859e
 size 4091