TechxGenus

/

starcoder2-3b-instruct

like 3

To fine-tune this model, you can directly use the original Alpaca script from [https://github.com/tatsu-lab/stanford_alpaca], with only the prompts and data modified. Configure it according to your specific hardware environment.

thank you for your effort.. i use the script [https://github.com/bigcode-project/starcoder2/blob/main/finetune.py] but i can't get the result as you .. is there some special action to the prompts ?

Use instruction fine-tuning scripts. That script is more like a continuous pre-training script.

thank you very much . as you said , i have done it successfully.