hello,i am honored to use your model but i met some problem when i load lora

#1
by thlw - opened

from peft import PeftModel,PeftConfig
finetune_model_path="/data/work/Llama2-Chinese/Llama2-Chinese-7b-Chat-LoRA"
config = PeftConfig.from_pretrained(finetune_model_path)
tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path,use_fast=False)
tokenizer.pad_token = tokenizer.eos_token
model = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path,device_map='auto',torch_dtype=torch.float16,load_in_8bit=True)
model = PeftModel.from_pretrained(model, finetune_model_path, device_map={"": 0})
model =model.eval()
input_ids = tokenizer(['Human: 介绍一下北京\nAssistant: '], return_tensors="pt",add_special_tokens=False).input_ids.to('cuda')
generate_input = {
"input_ids":input_ids,
"max_new_tokens":512,
"do_sample":True,
"top_k":50,
"top_p":0.95,
"temperature":0.3,
"repetition_penalty":1.3,
"eos_token_id":tokenizer.eos_token_id,
"bos_token_id":tokenizer.bos_token_id,
"pad_token_id":tokenizer.pad_token_id
}
generate_ids = model.generate(**generate_input)
text = tokenizer.decode(generate_ids[0])
print(text)

i just load the lora as description , but

image.png
i wander whether my llama2 is expired or other questions

LlamaFamily org

use peft to load

Sign up or log in to comment