search

unslothai / unsloth

Finetune Llama 3.1, Mistral, Phi & Gemma LLMs 2-5x faster with 80% less memory
https://unsloth.ai
Apache License 2.0
16.22k stars 1.11k forks source link

RuntimeError: Unsloth: `unsloth/llama-3-8b-bnb-4bit` is not a full model or a PEFT model. #614

Open Lipapaldl opened 3 months ago

Lipapaldl commented 3 months ago

RuntimeError: Unsloth: unsloth/llama-3-8b-bnb-4bit is not a full model or a PEFT model.

danielhanchen commented 3 months ago

Are you connected to the internet?

rogerslh commented 3 months ago

RuntimeError: Unsloth：unsloth/llama-3-8b-bnb-4bit不是完整模型或 PEFT 模型。

I have encountered the same problem. Have you resolved it?

rogerslh commented 3 months ago

Are you connected to the internet?

sure

danielhanchen commented 2 months ago

Normally if Colab works, then it should be fine - did you all clone the colab notebook exactly?

shafferjohn commented 2 months ago

It maybe something wrong with your network or proxy

The following is my solution, it works for me.

Download this model mannually: git lfs clone https://huggingface.co/unsloth/llama-3-8b-bnb-4bit

Change model_name to your local path model_name = "unsloth/llama-3-8b-bnb-4bit" ==> model_name = "./llama-3-8b-bnb-4bit"

Let me know if it works :)

zhuangkun5 commented 4 days ago

It maybe something wrong with your network or proxy

The following is my solution, it works for me.

Download this model mannually: git lfs clone https://huggingface.co/unsloth/llama-3-8b-bnb-4bit

Change model_name to your local path model_name = "unsloth/llama-3-8b-bnb-4bit" ==> model_name = "./llama-3-8b-bnb-4bit"

Let me know if it works :) 大佬能不能具体一点，我的llama模型的位置是："F:\unsloth-win11\model\llama-3-8b-bnb-4bit"我该如何设置model_name