🚀 LongWriter-llama3.1-8b

LongWriter-llama3.1-8b 是一个基于 Meta-Llama-3.1-8B 训练的模型，能够一次性生成 10000 字以上的内容。

📚 [LongWriter 数据集] • 💻 [GitHub 仓库] • 📃 [LongWriter 论文]

🚀 快速开始

环境要求

运行该模型需要 transformers>=4.43.0 的环境。

提示模板

请遵循以下提示模板（系统提示可选）：<<SYS>>\n{系统提示}\n<</SYS>>\n\n[INST]{查询1}[/INST]{回复1}[INST]{查询2}[/INST]{回复2}...

💻 使用示例

基础用法

以下是一个简单的模型部署示例：

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
tokenizer = AutoTokenizer.from_pretrained("THUDM/LongWriter-llama3.1-8b", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("THUDM/LongWriter-llama3.1-8b", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto")
model = model.eval()
query = "Write a 10000-word China travel guide"
prompt = f"[INST]{query}[/INST]"
input = tokenizer(prompt, truncation=False, return_tensors="pt").to(device)
context_length = input.input_ids.shape[-1]
output = model.generate(
    **input,
    max_new_tokens=32768,
    num_beams=1,
    do_sample=True,
    temperature=0.5,
)[0]
response = tokenizer.decode(output[context_length:], skip_special_tokens=True)
print(response)

高级用法

你也可以使用 vllm 部署该模型，它可以在一分钟内生成 10000 字以上的内容。以下是示例代码：

model = LLM(
    model= "THUDM/LongWriter-llama3.1-8b",
    dtype="auto",
    trust_remote_code=True,
    tensor_parallel_size=1,
    max_model_len=32768,
    gpu_memory_utilization=0.5,
)
tokenizer = model.get_tokenizer()
generation_params = SamplingParams(
    temperature=0.5,
    top_p=0.8,
    top_k=50,
    max_tokens=32768,
    repetition_penalty=1,
)
query = "Write a 10000-word China travel guide"
prompt = f"[INST]{query}[/INST]"
input_ids = tokenizer(prompt, truncation=False, return_tensors="pt").input_ids[0].tolist()
outputs = model.generate(
    sampling_params=generation_params,
    prompt_token_ids=[input_ids],
)
output = outputs[0]
print(output.outputs[0].text)

📄 许可证

该模型遵循 Llama-3.1 许可证。

📚 引用

如果您觉得我们的工作有用，请考虑引用 LongWriter：

@article{bai2024longwriter,
  title={LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMs}, 
  author={Yushi Bai and Jiajie Zhang and Xin Lv and Linzhi Zheng and Siqi Zhu and Lei Hou and Yuxiao Dong and Jie Tang and Juanzi Li},
  journal={arXiv preprint arXiv:2408.07055},
  year={2024}
}

📦 相关信息