Training in progress, step 500, checkpoint

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bf0f1647148ad1812b80ad783b74992c9e0c032d6b7e3e74894c4f908872092e
-size 1474893573
+oid sha256:31690aea3905a53fd850b72ab4b0994be40bd3b2a107646d7eb89979303cab2f
+size 1474893317

checkpoint-500/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:99c0e8eea89a5b5db3ec672dd872480cdebf1869769e7cc02f6d262e5910910b
+oid sha256:e6baa0ed3c0013d58345b0e3c81e7ffae409f63038ae73b3a5056a229c83bff9
 size 737457141

checkpoint-500/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0a01fce91b7ff51ddf06ce50055bf968e88f4a84a7cff6b7315179b09ce1f6b6
+oid sha256:60299c0baea1d4674e08a12c849495bf25d6fb27475f0e1217c4c12270589427
 size 14575

checkpoint-500/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:63f9421e0151b550d7e4cd1fb14a829d5928b2942e35b57c5b2fe0a1ffa12b17
+oid sha256:6f5807a4971dd52ca29e62bb6fe4ab951bf295105e514fbf391e7c26f976c560
 size 627

checkpoint-500/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": 1.0330578512396693,
   "eval_steps": 50,
   "global_step": 500,
@@ -10,106 +10,166 @@
   "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
     },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.0002985365853658536,
+      "loss": 0.6744,
+      "step": 350
+    },
     {
       "epoch": 0.72,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5931060910224915,
-      "eval_runtime": 29.0233,
-      "eval_samples_per_second": 40.795,
-      "eval_steps_per_second": 20.397,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.5975,
+      "eval_samples_per_second": 38.696,
+      "eval_steps_per_second": 19.348,
       "step": 350
     },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.0002802439024390244,
+      "loss": 0.6448,
+      "step": 400
+    },
     {
       "epoch": 0.83,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5643754005432129,
-      "eval_runtime": 28.8155,
-      "eval_samples_per_second": 41.089,
-      "eval_steps_per_second": 20.545,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.6461,
+      "eval_samples_per_second": 38.635,
+      "eval_steps_per_second": 19.317,
       "step": 400
     },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.0002619512195121951,
+      "loss": 0.7365,
+      "step": 450
+    },
     {
       "epoch": 0.93,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5633080005645752,
-      "eval_runtime": 29.0501,
-      "eval_samples_per_second": 40.757,
-      "eval_steps_per_second": 20.379,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.713,
+      "eval_samples_per_second": 38.55,
+      "eval_steps_per_second": 19.275,
       "step": 450
     },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00024365853658536585,
+      "loss": 0.6083,
+      "step": 500
+    },
     {
       "epoch": 1.03,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5926231145858765,
-      "eval_runtime": 28.7918,
-      "eval_samples_per_second": 41.123,
-      "eval_steps_per_second": 20.561,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.6783,
+      "eval_samples_per_second": 38.594,
+      "eval_steps_per_second": 19.297,
       "step": 500
     }
   ],
-  "logging_steps": 1000,
+  "logging_steps": 50,
   "max_steps": 1000,
   "num_train_epochs": 3,
   "save_steps": 100,
-  "total_flos": 1611348750173832.0,
+  "total_flos": 1742906637149832.0,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-500/training_args.bin

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