Qwen1.5-14B-Chat使用与多方式部署（Linux和GPU环境）

下载模型权重

使用modelscope下载

from modelscope.models import Model
model = Model.from_pretrained('qwen/Qwen1.5-14B-chat')

使用huggingface下载

from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained('Qwen/Qwen1.5-14B-Chat')

用Transformers使用模型

from transformers import AutoModelForCausalLM, AutoTokenizer
device = "cuda" # the device to load the model onto

model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen1.5-14B-Chat",
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen1.5-14B-Chat")

prompt = "Give me a short introduction to large language model."
messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(device)

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]

vLLM 部署模型

vLLM拉起openai服务

python -m vllm.entrypoints.openai.api_server \
    --model qwen/Qwen1.5-14B-Chat --max-model-len 8192  --gpu-memory-utilization 0.95

访问服务

curl http://localhost:8000/v1/chat/completions \
    -H "Content-Type: application/json" \
    -d '{
        "model": "qwen/Qwen1.5-7B-Chat",
        "messages": [
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": "写一篇春天为主题的作文"}
        ],
        "stop": ["<|im_end|>", "<|endoftext|>"]
    }'

参考资料：

• https://huggingface.co/Qwen/Qwen1.5-14B-Chat
• https://developer.aliyun.com/article/1439006