Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
|
@@ -3,6 +3,10 @@ from vllm import LLM, SamplingParams
|
|
| 3 |
from pydantic import BaseModel
|
| 4 |
from fastapi.responses import JSONResponse
|
| 5 |
import uvicorn
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
|
| 7 |
# 1. 加载模型(会自动从 HuggingFace Hub 拉取)
|
| 8 |
llm = LLM(model="Qwen/Qwen2.5-Coder-7B-Instruct")
|
|
|
|
| 3 |
from pydantic import BaseModel
|
| 4 |
from fastapi.responses import JSONResponse
|
| 5 |
import uvicorn
|
| 6 |
+
import os
|
| 7 |
+
os.environ["HF_HOME"] = "/data"
|
| 8 |
+
os.environ["TRANSFORMERS_CACHE"] = "/data"
|
| 9 |
+
os.environ["HF_HUB_CACHE"] = "/data"
|
| 10 |
|
| 11 |
# 1. 加载模型(会自动从 HuggingFace Hub 拉取)
|
| 12 |
llm = LLM(model="Qwen/Qwen2.5-Coder-7B-Instruct")
|