search

unslothai / unsloth

Finetune Llama 3.2, Mistral, Phi, Qwen 2.5 & Gemma LLMs 2-5x faster with 80% less memory
https://unsloth.ai
Apache License 2.0
18.37k stars 1.28k forks source link

Pythia Models Unsupported #1306

Open sert121 opened 3 days ago

sert121 commented 3 days ago

ModelConfig(model_name_or_path='EleutherAI/pythia-2.8b', model_revision='main', torch_dtype=None, trust_remote_code=False, attn_implementation=None, use_peft=True, lora_r=32, lora_alpha=16, lora_dropout=0.2, lora_target_modules=None, lora_modules_to_save=None, lora_task_type='CAUSAL_LM', load_in_8bit=False, load_in_4bit=False, bnb_4bit_quant_type='nf4', use_bnb_nested_quant=False) Traceback (most recent call last): raise NotImplementedError( NotImplementedError: Unsloth: EleutherAI/pythia-2.8b not supported yet! Make an issue to https://github.com/unslothai/unsloth!

danielhanchen commented 7 hours ago

All model support which will come in a future update will support it!! We're working on it!