使用 Cerebrium 部署#
vLLM 可以运行在基于云的 GPU 机器上,使用 Cerebrium,这是一个无服务器 AI 基础设施平台,使公司更容易构建和部署基于 AI 的应用程序。
要安装 Cerebrium 客户端,请运行:
$ pip install cerebrium
$ cerebrium login
接下来,创建你的 Cerebrium 项目,运行:
$ cerebrium init vllm-project
接下来,要安装所需的软件包,请将以下内容添加到你的 cerebrium.toml 中:
[cerebrium.deployment]
docker_base_image_url = "nvidia/cuda:12.1.1-runtime-ubuntu22.04"
[cerebrium.dependencies.pip]
vllm = "latest"
接下来,让我们添加代码来处理你选择的 LLM 的推理(本例中为 mistralai/Mistral-7B-Instruct-v0.1),将以下代码添加到你的 main.py 中:
from vllm import LLM, SamplingParams
llm = LLM(model="mistralai/Mistral-7B-Instruct-v0.1")
def run(prompts: list[str], temperature: float = 0.8, top_p: float = 0.95):
sampling_params = SamplingParams(temperature=temperature, top_p=top_p)
outputs = llm.generate(prompts, sampling_params)
# Print the outputs.
results = []
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
results.append({"prompt": prompt, "generated_text": generated_text})
return {"results": results}
然后,运行以下代码将其部署到云中
$ cerebrium deploy
如果成功,你应该会收到一个 CURL 命令,你可以使用它来调用推理。只需记住在 URL 末尾添加你调用的函数名称(在本例中为 /run)
curl -X POST https://api.cortex.cerebrium.ai/v4/p-xxxxxx/vllm/run \
-H 'Content-Type: application/json' \
-H 'Authorization: <JWT TOKEN>' \
--data '{
"prompts": [
"Hello, my name is",
"The president of the United States is",
"The capital of France is",
"The future of AI is"
]
}'
你应该会收到类似以下的响应:
{
"run_id": "52911756-3066-9ae8-bcc9-d9129d1bd262",
"result": {
"result": [
{
"prompt": "Hello, my name is",
"generated_text": " Sarah, and I'm a teacher. I teach elementary school students. One of"
},
{
"prompt": "The president of the United States is",
"generated_text": " elected every four years. This is a democratic system.\n\n5. What"
},
{
"prompt": "The capital of France is",
"generated_text": " Paris.\n"
},
{
"prompt": "The future of AI is",
"generated_text": " bright, but it's important to approach it with a balanced and nuanced perspective."
}
]
},
"run_time_ms": 152.53663063049316
}
你现在拥有一个自动扩展的端点,你只需为使用的计算付费!