#Multi-GPU possible for SFT with Unsloth?

2 messages · Page 1 of 1 (latest)

west carbon
#

Hi! I am doing LoRA SFT with Unsloth and was wondering if it was possible to train on multiple GPUs ? If yes, what parameter should I modify to allow it ? Thanks !

trail trench
#

https://docs.unsloth.ai/basics/multi-gpu-training-with-unsloth

Personally, I also had to do this: https://github.com/artidoro/qlora/issues/186#issuecomment-1943045599

Multi-GPU Training with Unsloth | Unsloth Documentation

Learn how to fine-tune LLMs on multiple GPUs and parallelism with Unsloth.

GitHub
ValueError: You can't train a model that has been loaded in 8-bit p...

I'm new to this field and want to try it out. I met this problem when trying to run the shell (finetune_guanaco_7b.sh) or "python qlora.py –learning_rate 0.0001 --model_name_or_path huggyl...