Training in progress, epoch 8

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e9efe4e663f94d9d0e461d2be685627b87d87f9c2a46bd0cef4eba7214e985d7
+oid sha256:763c0f2f7e203bb5429b6ab48b8a91d01a53d286ebff1b61e3f1b5891f786026
 size 267832560

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

@@ -38,3 +38,5 @@ You should probably TRAIN this model on a down-stream task to be able to use it
 {'eval_loss': 0.31772053241729736, 'eval_accuracy': 0.87, 'eval_runtime': 37.1806, 'eval_samples_per_second': 26.896, 'eval_steps_per_second': 0.861, 'epoch': 4.0}
 {'eval_loss': 0.2808445990085602, 'eval_accuracy': 0.932, 'eval_runtime': 37.3397, 'eval_samples_per_second': 26.781, 'eval_steps_per_second': 0.857, 'epoch': 5.0}
 {'eval_loss': 0.3926897644996643, 'eval_accuracy': 0.905, 'eval_runtime': 37.6368, 'eval_samples_per_second': 26.57, 'eval_steps_per_second': 0.85, 'epoch': 6.0}
+{'eval_loss': 0.37185582518577576, 'eval_accuracy': 0.922, 'eval_runtime': 37.484, 'eval_samples_per_second': 26.678, 'eval_steps_per_second': 0.854, 'epoch': 7.0}
+{'loss': 0.1013, 'grad_norm': 0.6478258371353149, 'learning_rate': 8.400000000000001e-06, 'epoch': 8.0}

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

@@ -1 +1 @@
-{"eval/loss": 0.37185582518577576, "eval/accuracy": 0.922, "eval/runtime": 37.484, "eval/samples_per_second": 26.678, "eval/steps_per_second": 0.854, "train/epoch": 7.0, "train/global_step": 875, "_timestamp": 1727232299.565779, "_runtime": 22708.692869901657, "_step": 15, "train/loss": 0.2956, "train/grad_norm": 2.695140838623047, "train/learning_rate": 9.200000000000002e-06, "train_runtime": 8026.8642, "train_samples_per_second": 2.492, "train_steps_per_second": 0.156, "total_flos": 2396475988298112.0, "train_loss": 0.11480112991333008}
+{"eval/loss": 0.4580109715461731, "eval/accuracy": 0.91, "eval/runtime": 38.2702, "eval/samples_per_second": 26.13, "eval/steps_per_second": 0.836, "train/epoch": 8.0, "train/global_step": 1000, "_timestamp": 1727233994.321086, "_runtime": 24403.44817686081, "_step": 17, "train/loss": 0.1013, "train/grad_norm": 0.6478258371353149, "train/learning_rate": 8.400000000000001e-06, "train_runtime": 8026.8642, "train_samples_per_second": 2.492, "train_steps_per_second": 0.156, "total_flos": 2396475988298112.0, "train_loss": 0.11480112991333008}