Training in progress, epoch 3

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fc2934c19e3f145b2c8af3a46df740321a95d0b384d33395bbea1d2f8f05e99a
+oid sha256:c7b41357485e1704d509aa0a1421fd007fb7be5a861dfd0e44630f9fa473d8a3
 size 409103316

run-0/checkpoint-1599/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:70e29b5f49710d0d17c6aee07e8cc1bc4edaa86a3c66b73d52d1ef1136d10f23
+oid sha256:c7b41357485e1704d509aa0a1421fd007fb7be5a861dfd0e44630f9fa473d8a3
 size 409103316

run-0/checkpoint-1599/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b04eae6a685408364150ca5d02a35737876cace088d1b38561f5da6ab19d7a30
+oid sha256:aaad4a07ff4c40b92bd14a50098833d4eb8b48978464d3b4574b5ac088e998eb
 size 818327802

run-0/checkpoint-1599/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b3a7f77912cd35b35f45f50fd2c3bbb7b15cbe8d92a9601be01ae387622324a8
+oid sha256:4424c5078b48aec91480ea9ddc98250169be3ffb90e1ca9a96b04048992a7ca6
 size 1064

run-0/checkpoint-1599/trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.6369415521621704,
-  "best_model_checkpoint": "BERT-WMM/run-0/checkpoint-533",
+  "best_metric": 0.75532066822052,
+  "best_model_checkpoint": "BERT-WMM/run-0/checkpoint-1599",
   "epoch": 3.0,
   "eval_steps": 500,
   "global_step": 1599,
@@ -10,65 +10,62 @@
   "log_history": [
     {
       "epoch": 0.94,
-      "grad_norm": 11.681629180908203,
-      "learning_rate": 2.5842410540171958e-05,
-      "loss": 0.7181,
+      "grad_norm": 10.492278099060059,
+      "learning_rate": 7.698774119953696e-07,
+      "loss": 0.9956,
       "step": 500
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.7525821596244131,
-      "eval_loss": 0.6369415521621704,
-      "eval_runtime": 1.9836,
-      "eval_samples_per_second": 1073.792,
-      "eval_steps_per_second": 67.553,
+      "eval_accuracy": 0.6281690140845071,
+      "eval_loss": 0.8787437081336975,
+      "eval_runtime": 2.0952,
+      "eval_samples_per_second": 1016.614,
+      "eval_steps_per_second": 63.956,
       "step": 533
     },
     {
       "epoch": 1.88,
-      "grad_norm": 6.492458820343018,
-      "learning_rate": 2.1053231418568556e-05,
-      "loss": 0.4571,
+      "grad_norm": 16.712677001953125,
+      "learning_rate": 4.1961471318036977e-07,
+      "loss": 0.858,
       "step": 1000
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.7497652582159624,
-      "eval_loss": 0.6769182682037354,
-      "eval_runtime": 2.0259,
-      "eval_samples_per_second": 1051.385,
-      "eval_steps_per_second": 66.144,
+      "eval_accuracy": 0.6732394366197183,
+      "eval_loss": 0.791917085647583,
+      "eval_runtime": 2.1109,
+      "eval_samples_per_second": 1009.044,
+      "eval_steps_per_second": 63.48,
       "step": 1066
     },
     {
       "epoch": 2.81,
-      "grad_norm": 13.743987083435059,
-      "learning_rate": 1.626405229696515e-05,
-      "loss": 0.2889,
+      "grad_norm": 29.26333999633789,
+      "learning_rate": 6.935201436536996e-08,
+      "loss": 0.7846,
       "step": 1500
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.7596244131455399,
-      "eval_loss": 0.8726277351379395,
-      "eval_runtime": 2.0636,
-      "eval_samples_per_second": 1032.174,
-      "eval_steps_per_second": 64.935,
+      "eval_accuracy": 0.6892018779342723,
+      "eval_loss": 0.75532066822052,
+      "eval_runtime": 2.0969,
+      "eval_samples_per_second": 1015.8,
+      "eval_steps_per_second": 63.905,
       "step": 1599
     }
   ],
   "logging_steps": 500,
-  "max_steps": 3198,
+  "max_steps": 1599,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 6,
+  "num_train_epochs": 3,
   "save_steps": 500,
   "total_flos": 507646505902536.0,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": {
-    "learning_rate": 3.063158966177536e-05,
-    "num_train_epochs": 6,
-    "per_device_train_batch_size": 16,
-    "weight_decay": 2.6426532465921284e-05
+    "learning_rate": 1.1201401108103694e-06
   }
 }

run-0/checkpoint-1599/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:48c24530a27c9d1e90a961cc0645745786b44285899342b0769165200ee8a2b1
+oid sha256:f3b32ee6920eea73a8ff94d684bfa22384952a503144760a3a67dfa2dfe3f39f
 size 4856

runs/Apr18_16-31-07_544fc269209b/events.out.tfevents.1713457870.544fc269209b.792.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8aea9294fd5b07cc0e83aa6ecc325232e8f25d9c6601fc8a5d6d9e88eee0356f
-size 5962
+oid sha256:22075c414dff056bcd45a0e4e13a3d05f01c9aa4df3e19fc19dd8b647078a313
+size 6850