Codellama Chat 13b Chinese

🚀 CodeLlaMa Chinese Version (Supporting Multi-round Conversations)

This is a Chinese version of the CodeLlaMa model, specifically designed for code assistance. It's suitable for answering code-related questions, different from the ChineseLlaMa model.

📚 Documentation

• Model Introduction: CodeLlaMa is a model dedicated to code assistance. Unlike ChineseLlaMa, it is more suitable for responding to code-related questions.
• Related GitHub Repository: CodeLLaMA-chat

💻 Usage Examples

Basic Usage

# 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()

📄 License

The model is released under the OpenRAIL license.

📋 Information Table