| # how to use | |
| need install "pip install git+https://github.com/cczhong11/AutoGPTQ" before https://github.com/PanQiWei/AutoGPTQ/pull/189 got merged | |
| ``` | |
| from transformers import AutoTokenizer, AutoModelForCausalLM | |
| from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig | |
| quantized_model_dir = "cczhong/internlm-chat-7b-4bit-gptq" | |
| tokenizer = AutoTokenizer.from_pretrained(quantized_model_dir, trust_remote_code=True) | |
| model = AutoGPTQForCausalLM.from_quantized(quantized_model_dir, device="cuda:0",trust_remote_code=True) | |
| response, history = model.chat(tokenizer, "你好", history=[]) | |
| ``` |