Training in progress, step 1

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3c2765fe658f05ee92acdfed61731ad331368977e42ca5301d0bf811a1862725
+oid sha256:2394f54f8f2300f62835eff3d6b84e2cc019ba8e84b34cff88d0bf9480a9e792
 size 155589344

train.log

@@ -1 +1 @@
-03/19/2025 15:38:07 - INFO - root -   args: Namespace(output_name='llama1b_lora', datasets=['harmful_behavior'], student_model='distilbert/distilgpt2', teacher_model='distilbert/distilgpt2', alpha=0.0, temperature=1.0, num_train_epochs=1, learning_rate=0.001, max_num_tokens=1024, batch_size=1, grad_acc_steps=16, weight_decay=0.01, adam_epsilon=1e-08, warmup_steps=0, max_grad_norm=1.0, dropout=0.1, kl_loss_weight=0, exclude_neg=True, no_weights=True, optim='adamw_torch', max_steps=1, lr_scheduler_type='linear', logging_steps=50, save_epochs=10, seed=2, resume_from_checkpoint=False, hub_strategy='all_checkpoints', report_to='none', push_to_hub=True, proportions=[1.0], split='train', streaming=True, sequence_length=512, model_dir='./trained/', is_lora_model=False, lora=True, r=16, lora_alpha=32, lora_dropout=0.1, task_type='CAUSAL_LM', output_dir='./trained/llama1b_lora', logger=<RootLogger root (INFO)>)
+03/19/2025 16:24:18 - INFO - root -   args: Namespace(output_name='llama1b_lora', datasets=['harmful_behavior'], student_model='distilbert/distilgpt2', teacher_model='distilbert/distilgpt2', alpha=0.0, temperature=1.0, num_train_epochs=1, learning_rate=0.001, max_num_tokens=1024, batch_size=1, grad_acc_steps=16, weight_decay=0.01, adam_epsilon=1e-08, warmup_steps=0, max_grad_norm=1.0, dropout=0.1, kl_loss_weight=0, exclude_neg=True, no_weights=True, optim='adamw_torch', max_steps=1, lr_scheduler_type='linear', logging_steps=50, save_epochs=10, seed=2, resume_from_checkpoint=False, hub_strategy='all_checkpoints', report_to='none', push_to_hub=True, proportions=[1.0], split='train', streaming=True, sequence_length=512, model_dir='./trained/', is_lora_model=False, lora=True, r=16, lora_alpha=32, lora_dropout=0.1, task_type='CAUSAL_LM', output_dir='./trained/llama1b_lora', logger=<RootLogger root (INFO)>)

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:588e764622c13e19bc66264d1acb25c3c0f5ccfbf1c2c3198f3c7de395f2b65c
+oid sha256:8d1b0461349b0644b9b60de9db49258b64bc27f12a88274a942313792e0db2f8
 size 5304