Qwen1.5

聊天

模型加载：

model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen1.5-0.5B-Chat",
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen1.5-0.5B-Chat")

Prompt：

prompt = "如何使用 Python 写一个 Hello World？"
messages = [
    {"role": "system", "content": "你是一个有用的助手。"},
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
print(text)
model_inputs = tokenizer([text], return_tensors="pt").to(device)

回答：

generated_ids = model.generate(
    model_inputs.input_ids,
    max_new_tokens=512
)
generated_ids = [
    output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(response)

网络资源

• Qwen1.5 介绍 | Qwen
• 通义千问Qwen-7B效果如何？Firefly微调实践，效果出色
• Qwen (Qwen) HuggingFace
• Fine-tuning and Quantization of Qwen1.5 LLMs on Your Computer
• Qwen
• ssbuild/qwen_finetuning: qwen-7b and qwen-14b finetuning

本文作者：Maeiee

本文链接：Qwen1.5

版权声明：如无特别声明，本文即为原创文章，版权归 Maeiee 所有，未经允许不得转载！

喜欢我文章的朋友请随缘打赏，鼓励我创作更多更好的作品！