Training in progress, epoch 4

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c8091a6d50c16ffaec776032560df90efcaf84d3ef650fe1893203fe2c3b07e
+oid sha256:9db619b5890ee02e32aa5b3b4461ed2b1be4ee2f6b5b4c3bf3c45d0e47385936
 size 267832560

wandb/run-20240924_172630-x9iddikd/files/output.log

@@ -22,3 +22,5 @@ You should probably TRAIN this model on a down-stream task to be able to use it
 {'eval_loss': 0.21834564208984375, 'eval_accuracy': 0.938, 'eval_runtime': 30.3683, 'eval_samples_per_second': 32.929, 'eval_steps_per_second': 2.075, 'epoch': 1.0}
 {'loss': 0.2031, 'grad_norm': 1.1480563879013062, 'learning_rate': 1.2e-05, 'epoch': 2.0}
 {'eval_loss': 0.19427122175693512, 'eval_accuracy': 0.938, 'eval_runtime': 42.2287, 'eval_samples_per_second': 23.681, 'eval_steps_per_second': 1.492, 'epoch': 2.0}
+{'eval_loss': 0.3195326626300812, 'eval_accuracy': 0.921, 'eval_runtime': 26.5577, 'eval_samples_per_second': 37.654, 'eval_steps_per_second': 2.372, 'epoch': 3.0}
+{'loss': 0.0672, 'grad_norm': 1.1029362678527832, 'learning_rate': 4.000000000000001e-06, 'epoch': 4.0}

wandb/run-20240924_172630-x9iddikd/files/wandb-summary.json

@@ -1 +1 @@
-{"eval/loss": 0.3195326626300812, "eval/accuracy": 0.921, "eval/runtime": 26.5577, "eval/samples_per_second": 37.654, "eval/steps_per_second": 2.372, "train/epoch": 3.0, "train/global_step": 750, "_timestamp": 1727214317.037596, "_runtime": 4726.164686918259, "_step": 3, "train/loss": 0.2031, "train/grad_norm": 1.1480563879013062, "train/learning_rate": 1.2e-05}
+{"eval/loss": 0.36123067140579224, "eval/accuracy": 0.925, "eval/runtime": 26.675, "eval/samples_per_second": 37.488, "eval/steps_per_second": 2.362, "train/epoch": 4.0, "train/global_step": 1000, "_timestamp": 1727215983.911635, "_runtime": 6393.038725852966, "_step": 5, "train/loss": 0.0672, "train/grad_norm": 1.1029362678527832, "train/learning_rate": 4.000000000000001e-06}