The model's quantization config from the arguments has no `quant_method` attribute

#1
by fanface - opened

when I use this way to load model

from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("mlx-community/Qwen2.5-7B-Instruct-1M-4bit")
model = AutoModelForCausalLM.from_pretrained("mlx-community/Qwen2.5-7B-Instruct-1M-4bit")

python comlain The model's quantization config from the arguments has no quant_method attribute

how to fix it? or should I just use mlx to interact with this model?

Sign up or log in to comment