Training in progress, step 900, checkpoint

checkpoint-900/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b7689cb9e578e6ff23c472b6b89d44ec2dc061eb80574d224770d15d48882a8b
-size 1474893573
+oid sha256:67cf86b99d0132fd320f16039886061bce32dd7dbbbed110bf32678d77202c7d
+size 1474893317

checkpoint-900/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:baa9ee16cd3de8ccdda427441b280c6d692e941eb8eaea43015b03066b176397
+oid sha256:e6baa0ed3c0013d58345b0e3c81e7ffae409f63038ae73b3a5056a229c83bff9
 size 737457141

checkpoint-900/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a1c1f569431d58283f3fbc997efce4b3e16cd01ee209c1f878c91243e40de338
+oid sha256:a9a3a4358d666810ee4fdec33dd73a01c1ff4b3535585eb7e750248e2d30a0ee
 size 14575

checkpoint-900/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dc356ec59d335c7b0c5cd452933492e17947e07adfe349a9499875c8d3aab306
+oid sha256:ae59fb5c7cc6520b33a32025d13d9814d580af05d2ce61e0493025b701c279a9
 size 627

checkpoint-900/trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.5623977184295654,
-  "best_model_checkpoint": "/kaggle/output/checkpoint-900",
+  "best_metric": 0.5696164965629578,
+  "best_model_checkpoint": "/kaggle/output/checkpoint-100",
   "epoch": 1.859504132231405,
   "eval_steps": 50,
   "global_step": 900,
@@ -10,178 +10,286 @@
   "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
     },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.00022536585365853657,
+      "loss": 0.6217,
+      "step": 550
+    },
     {
       "epoch": 1.14,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5648908019065857,
-      "eval_runtime": 28.8558,
-      "eval_samples_per_second": 41.032,
-      "eval_steps_per_second": 20.516,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.5897,
+      "eval_samples_per_second": 38.706,
+      "eval_steps_per_second": 19.353,
       "step": 550
     },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.00020707317073170728,
+      "loss": 0.642,
+      "step": 600
+    },
     {
       "epoch": 1.24,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5628103017807007,
-      "eval_runtime": 28.9143,
-      "eval_samples_per_second": 40.949,
-      "eval_steps_per_second": 20.474,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.6541,
+      "eval_samples_per_second": 38.624,
+      "eval_steps_per_second": 19.312,
       "step": 600
     },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018878048780487803,
+      "loss": 0.6433,
+      "step": 650
+    },
     {
       "epoch": 1.34,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5630568265914917,
-      "eval_runtime": 28.8676,
-      "eval_samples_per_second": 41.015,
-      "eval_steps_per_second": 20.507,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.5944,
+      "eval_samples_per_second": 38.7,
+      "eval_steps_per_second": 19.35,
       "step": 650
     },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.00017048780487804874,
+      "loss": 0.7497,
+      "step": 700
+    },
     {
       "epoch": 1.45,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5687929391860962,
-      "eval_runtime": 28.9082,
-      "eval_samples_per_second": 40.957,
-      "eval_steps_per_second": 20.479,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 31.0188,
+      "eval_samples_per_second": 38.17,
+      "eval_steps_per_second": 19.085,
       "step": 700
     },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.0001521951219512195,
+      "loss": 0.6385,
+      "step": 750
+    },
     {
       "epoch": 1.55,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5623889565467834,
-      "eval_runtime": 28.8406,
-      "eval_samples_per_second": 41.053,
-      "eval_steps_per_second": 20.527,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.6101,
+      "eval_samples_per_second": 38.68,
+      "eval_steps_per_second": 19.34,
       "step": 750
     },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00013390243902439023,
+      "loss": 0.6581,
+      "step": 800
+    },
     {
       "epoch": 1.65,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5629884600639343,
-      "eval_runtime": 28.8402,
-      "eval_samples_per_second": 41.054,
-      "eval_steps_per_second": 20.527,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.6994,
+      "eval_samples_per_second": 38.568,
+      "eval_steps_per_second": 19.284,
       "step": 800
     },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00011560975609756097,
+      "loss": 0.6201,
+      "step": 850
+    },
     {
       "epoch": 1.76,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5627905130386353,
-      "eval_runtime": 29.4658,
-      "eval_samples_per_second": 40.182,
-      "eval_steps_per_second": 20.091,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.6306,
+      "eval_samples_per_second": 38.654,
+      "eval_steps_per_second": 19.327,
       "step": 850
     },
+    {
+      "epoch": 1.86,
+      "learning_rate": 9.73170731707317e-05,
+      "loss": 0.6424,
+      "step": 900
+    },
     {
       "epoch": 1.86,
       "eval_accuracy": 0.75,
-      "eval_loss": 0.5623977184295654,
-      "eval_runtime": 29.0192,
-      "eval_samples_per_second": 40.801,
-      "eval_steps_per_second": 20.4,
+      "eval_loss": 0.5931239724159241,
+      "eval_runtime": 30.5992,
+      "eval_samples_per_second": 38.694,
+      "eval_steps_per_second": 19.347,
       "step": 900
     }
   ],
-  "logging_steps": 1000,
+  "logging_steps": 50,
   "max_steps": 1000,
   "num_train_epochs": 3,
   "save_steps": 100,
-  "total_flos": 2907702695903472.0,
+  "total_flos": 3144506892460272.0,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-900/training_args.bin

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