SusGen-GPT-Mistral
Collection
4 items • Updated
How to use WHATX/30k-Mistral-7B-v0.3-small with PEFT:
from peft import PeftModel
from transformers import AutoModelForCausalLM
base_model = AutoModelForCausalLM.from_pretrained("../ckpts/Mistral-7B-v0.3")
model = PeftModel.from_pretrained(base_model, "WHATX/30k-Mistral-7B-v0.3-small")The best checkpoint is 180-epoch.