How to push or shere lora adapter to hugging face hub?

🀗Transformers
1

hi, i trained falcon model and already set push_to_hub paramter in training argument, but they not working.

from transformers import TrainingArguments

output_dir = "chatb_f"
per_device_train_batch_size = 4
gradient_accumulation_steps = 4
optim = "paged_adamw_32bit"
save_steps = 60
logging_steps = 10
learning_rate = 2e-4
max_grad_norm = 0.3
max_steps = 60
warmup_ratio = 0.03
lr_scheduler_type = "constant"

training_arguments = TrainingArguments(
    output_dir=output_dir,
    per_device_train_batch_size=per_device_train_batch_size,
    gradient_accumulation_steps=gradient_accumulation_steps,
    optim=optim,
    save_steps=save_steps,
    logging_steps=logging_steps,
    learning_rate=learning_rate,
    fp16=True,
    max_grad_norm=max_grad_norm,
    max_steps=max_steps,
    warmup_ratio=warmup_ratio,
    group_by_length=True,
    lr_scheduler_type=lr_scheduler_type,
    push_to_hub = True
)


from trl import SFTTrainer

max_seq_length = 512

trainer = SFTTrainer(
    model=model,
    train_dataset=dataset,
    peft_config=peft_config,
    dataset_text_field="text",
    max_seq_length=max_seq_length,
    tokenizer=tokenizer,
    args=training_arguments,
)
2

Try trainer.model.push_to_hub(repo_name)

After training of course. The other suggestion is exploring what functions you have available to you with the peft model

1 Like