CodeLlaMa模型的中文化版本 （支持多轮对话）
科普：CodeLlaMa是专门用于代码助手的，与ChineseLlaMa不同，适用于代码类问题的回复。
用于多轮对话的推理代码：
（可以直接复制运行，默认会自动拉取该模型权重）
关联Github仓库：https://github.com/CrazyBoyM/CodeLLaMA-chat
# from Firefly
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
def main():
model_name = 'shareAI/CodeLLaMA-chat-13b-Chinese'
device = 'cuda'
max_new_tokens = 500 # 每轮对话最多生成多少个token
history_max_len = 1000 # 模型记忆的最大token长度
top_p = 0.9
temperature = 0.35
repetition_penalty = 1.0
model = AutoModelForCausalLM.from_pretrained(
model_name,
trust_remote_code=True,
low_cpu_mem_usage=True,
torch_dtype=torch.float16,
device_map='auto'
).to(device).eval()
tokenizer = AutoTokenizer.from_pretrained(
model_name,
trust_remote_code=True,
use_fast=False
)
history_token_ids = torch.tensor([[]], dtype=torch.long)
user_input = input('User：')
while True:
input_ids = tokenizer(user_input, return_tensors="pt", add_special_tokens=False).input_ids
eos_token_id = torch.tensor([[tokenizer.eos_token_id]], dtype=torch.long)
user_input_ids = torch.concat([input_ids, eos_token_id], dim=1)
history_token_ids = torch.concat((history_token_ids, user_input_ids), dim=1)
model_input_ids = history_token_ids[:, -history_max_len:].to(device)
with torch.no_grad():
outputs = model.generate(
input_ids=model_input_ids, max_new_tokens=max_new_tokens, do_sample=True, top_p=top_p,
temperature=temperature, repetition_penalty=repetition_penalty, eos_token_id=tokenizer.eos_token_id
)
model_input_ids_len = model_input_ids.size(1)
response_ids = outputs[:, model_input_ids_len:]
history_token_ids = torch.concat((history_token_ids, response_ids.cpu()), dim=1)
response = tokenizer.batch_decode(response_ids)
print("Bot：" + response[0].strip().replace(tokenizer.eos_token, ""))
user_input = input('User：')
if __name__ == '__main__':
main()