Training in progress, epoch 4

logs/events.out.tfevents.1719305887.852b1e905a9a.223.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ccce78a6de7b7ead06a2eaeb7095288a4e8dd2066b568e59b6f75294e9e29337
-size 6498
+oid sha256:b2ac4d7fa81b6904905124ea42da5585b5330205bcf5bbe14f7557aa4d2fc4eb
+size 7032

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cc1ea81d68885299a8a06cce03ef9b210fe932d9c8ee4477cb9e366ba6369340
+oid sha256:afb0fc63aa3b163fe161f0fc2e80b005f0c467955700d8a1e79b85ea43d3e170
 size 17549312

run-0/checkpoint-2108/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52e23161d13c2c0074293d23c46bb546af70686a45a81739153b9141d5aee932
+oid sha256:afb0fc63aa3b163fe161f0fc2e80b005f0c467955700d8a1e79b85ea43d3e170
 size 17549312

run-0/checkpoint-2108/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b231c0660c8389832990d020279e7e6bed45f149f650c87707b64d90e7e759f
+oid sha256:03bfd9340b47f91e67617eca810211baa7689f51a4f5097ecd60622492ca1ec0
 size 35123898

run-0/checkpoint-2108/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5cff89e4666999e97f90d1645b64057419c7dc0028c34edc93cf14e1576ac0da
+oid sha256:d78463cfbd46cd542a69701cf2c4f2244fad66669f0025f3033d82ed9ee45fb0
 size 1064

run-0/checkpoint-2108/trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.8256880733944955,
+  "best_metric": 0.7901376146788991,
   "best_model_checkpoint": "tiny-bert-sst2-distilled/run-0/checkpoint-2108",
   "epoch": 4.0,
   "eval_steps": 500,
@@ -10,73 +10,73 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 16.55626678466797,
-      "learning_rate": 5.1383294230414005e-05,
-      "loss": 1.5907,
+      "grad_norm": 5.95961856842041,
+      "learning_rate": 9.55389368279823e-06,
+      "loss": 1.5369,
       "step": 527
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.8061926605504587,
-      "eval_loss": 1.2464169263839722,
-      "eval_runtime": 2.4007,
-      "eval_samples_per_second": 363.232,
-      "eval_steps_per_second": 2.916,
+      "eval_accuracy": 0.7339449541284404,
+      "eval_loss": 1.2773902416229248,
+      "eval_runtime": 2.5957,
+      "eval_samples_per_second": 335.939,
+      "eval_steps_per_second": 2.697,
       "step": 527
     },
     {
       "epoch": 2.0,
-      "grad_norm": Infinity,
-      "learning_rate": 4.405675241279466e-05,
-      "loss": 0.9038,
+      "grad_norm": 15.252978324890137,
+      "learning_rate": 8.492349940265094e-06,
+      "loss": 1.2159,
       "step": 1054
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8107798165137615,
-      "eval_loss": 1.1235342025756836,
-      "eval_runtime": 2.3533,
-      "eval_samples_per_second": 370.54,
-      "eval_steps_per_second": 2.975,
+      "eval_accuracy": 0.7786697247706422,
+      "eval_loss": 1.022659182548523,
+      "eval_runtime": 2.5741,
+      "eval_samples_per_second": 338.753,
+      "eval_steps_per_second": 2.719,
       "step": 1054
     },
     {
       "epoch": 3.0,
-      "grad_norm": 48.55740737915039,
-      "learning_rate": 3.67162818084498e-05,
-      "loss": 0.6946,
+      "grad_norm": 18.01114845275879,
+      "learning_rate": 7.430806197731956e-06,
+      "loss": 1.0132,
       "step": 1581
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8176605504587156,
-      "eval_loss": 1.1027159690856934,
-      "eval_runtime": 2.3719,
-      "eval_samples_per_second": 367.637,
-      "eval_steps_per_second": 2.951,
+      "eval_accuracy": 0.7844036697247706,
+      "eval_loss": 0.9622328281402588,
+      "eval_runtime": 2.5865,
+      "eval_samples_per_second": 337.137,
+      "eval_steps_per_second": 2.706,
       "step": 1581
     },
     {
       "epoch": 4.0,
-      "grad_norm": 16.267213821411133,
-      "learning_rate": 2.9375811204104943e-05,
-      "loss": 0.595,
+      "grad_norm": 15.912079811096191,
+      "learning_rate": 6.371276769700781e-06,
+      "loss": 0.9206,
       "step": 2108
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8256880733944955,
-      "eval_loss": 1.0700539350509644,
-      "eval_runtime": 2.3638,
-      "eval_samples_per_second": 368.893,
-      "eval_steps_per_second": 2.961,
+      "eval_accuracy": 0.7901376146788991,
+      "eval_loss": 0.9278557896614075,
+      "eval_runtime": 2.5808,
+      "eval_samples_per_second": 337.878,
+      "eval_steps_per_second": 2.712,
       "step": 2108
     }
   ],
   "logging_steps": 500,
-  "max_steps": 4216,
+  "max_steps": 5270,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 8,
+  "num_train_epochs": 10,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -94,9 +94,9 @@
   "train_batch_size": 128,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.5167874928728581,
-    "learning_rate": 5.872376483475886e-05,
-    "num_train_epochs": 8,
-    "temperature": 5
+    "alpha": 0.27608826195592573,
+    "learning_rate": 1.0615437425331367e-05,
+    "num_train_epochs": 10,
+    "temperature": 2
   }
 }

run-0/checkpoint-2108/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f28149fe21091b257234d7cbe1611ee6ca88e3a7cef675e40e6d90410e6fc1a6
+oid sha256:c8b6a60f7b85b38fa45cddf1a417ee51250fe5822237403416bf2406ff2cdb84
 size 5176