Training in progress, step 300, checkpoint

checkpoint-300/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b6150d1c135a738dac4d25b66e75b236d64c3f0525b12465041655f1593d44c2
-size 1474893573
+oid sha256:e261a8240f9ba007a28a67f46f5463f2d5850e2f5b3c08fc10ba3787cf10911e
+size 1474893317

checkpoint-300/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a377fab70c74d3ad7f842b0b0fc2c605c6d11eccd4e7dc7aeb0bc2bf07586b71
+oid sha256:e6baa0ed3c0013d58345b0e3c81e7ffae409f63038ae73b3a5056a229c83bff9
 size 737457141

checkpoint-300/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:45ebcf8a06b0a47db4ba6cab824c8e6297fe6600da3ed012c2d56464f795d894
+oid sha256:c738bf680db38eee6ca69048f2fb3d7738943fed94de94de434718429e42b80b
 size 14575

checkpoint-300/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8ef0b4df7167be3cee4946a24596e5af484f71a106b786dd8b9e0ad46b5a51b4
+oid sha256:738b3803967108b90cf940ac65ab8d2d54fdd8d26b5509e734c1ef7e40612c3c
 size 627

checkpoint-300/trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.5632852911949158,
-  "best_model_checkpoint": "/kaggle/output/checkpoint-300",
+  "best_metric": 0.5696164965629578,
+  "best_model_checkpoint": "/kaggle/output/checkpoint-100",
   "epoch": 0.6198347107438017,
   "eval_steps": 50,
   "global_step": 300,
@@ -10,70 +10,106 @@
   "log_history": [
     {
       "epoch": 0.0,
-      "learning_rate": 1.6666666666666667e-06,
-      "loss": 0.6979,
+      "learning_rate": 0.0,
+      "loss": 0.7246,
       "step": 1
     },
+    {
+      "epoch": 0.1,
+      "learning_rate": 7.333333333333332e-05,
+      "loss": 0.5982,
+      "step": 50
+    },
     {
       "epoch": 0.1,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.6723740696907043,
-      "eval_runtime": 28.903,
-      "eval_samples_per_second": 40.965,
-      "eval_steps_per_second": 20.482,
+      "eval_loss": 0.6296697854995728,
+      "eval_runtime": 30.5612,
+      "eval_samples_per_second": 38.742,
+      "eval_steps_per_second": 19.371,
       "step": 50
     },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00015,
+      "loss": 0.5505,
+      "step": 100
+    },
     {
       "epoch": 0.21,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.6272028088569641,
-      "eval_runtime": 28.756,
-      "eval_samples_per_second": 41.174,
-      "eval_steps_per_second": 20.587,
+      "eval_loss": 0.5696164965629578,
+      "eval_runtime": 30.5816,
+      "eval_samples_per_second": 38.716,
+      "eval_steps_per_second": 19.358,
       "step": 100
     },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0002333333333333333,
+      "loss": 0.5838,
+      "step": 150
+    },
     {
       "epoch": 0.31,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5726658701896667,
-      "eval_runtime": 28.7791,
-      "eval_samples_per_second": 41.141,
-      "eval_steps_per_second": 20.57,
+      "eval_loss": 0.562861442565918,
+      "eval_runtime": 31.0771,
+      "eval_samples_per_second": 38.099,
+      "eval_steps_per_second": 19.049,
       "step": 150
     },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00025333333333333333,
+      "loss": 0.5925,
+      "step": 200
+    },
     {
       "epoch": 0.41,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5658872723579407,
-      "eval_runtime": 28.8005,
-      "eval_samples_per_second": 41.11,
-      "eval_steps_per_second": 20.555,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.6023,
+      "eval_samples_per_second": 38.69,
+      "eval_steps_per_second": 19.345,
       "step": 200
     },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00025333333333333333,
+      "loss": 0.7003,
+      "step": 250
+    },
     {
       "epoch": 0.52,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5675005316734314,
-      "eval_runtime": 28.7479,
-      "eval_samples_per_second": 41.186,
-      "eval_steps_per_second": 20.593,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.7477,
+      "eval_samples_per_second": 38.507,
+      "eval_steps_per_second": 19.253,
       "step": 250
     },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00025333333333333333,
+      "loss": 0.606,
+      "step": 300
+    },
     {
       "epoch": 0.62,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5632852911949158,
-      "eval_runtime": 28.807,
-      "eval_samples_per_second": 41.101,
-      "eval_steps_per_second": 20.551,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.6306,
+      "eval_samples_per_second": 38.654,
+      "eval_steps_per_second": 19.327,
       "step": 300
     }
   ],
-  "logging_steps": 1000,
+  "logging_steps": 50,
   "max_steps": 1000,
   "num_train_epochs": 3,
   "save_steps": 100,
-  "total_flos": 964310061370152.0,
+  "total_flos": 1043244793555752.0,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-300/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fbe5e88ca49040823bdb83dc035b8a4470c58ae105b0482f624b7d48f5e08cbe
+oid sha256:cd0372335b8d69875fb40a5865f6914bab9ff539640d01b439e8f707b1e5a43d
 size 4027