Update Dockerfile

Dockerfile

@@ -21,5 +21,5 @@ RUN pip install --no-cache-dir --upgrade pip && \
 # CMD ["TRANSFORMERS_CACHE=./", "infinity_emb", "--model-name-or-path", "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2", "--port", "7860"]
 # CMD ["python", "app.py"]
 
-# CMD ["sh", "start-m3-server.sh"]
-CMD ["sh", "-c", "HF_HOME=/tmp/cache", "python", "m3_server.py"]
+CMD ["sh", "start-m3-server.sh"]
+# CMD ["sh", "-c", "HF_HOME=/tmp/cache", "python", "m3_server.py"]

m3_server.py

@@ -1,6 +1,7 @@
 import asyncio
 import os
 import time
+from pathlib import Path
 from concurrent.futures import ThreadPoolExecutor
 from typing import List, Tuple, Union
 from uuid import uuid4
@@ -11,7 +12,8 @@ from FlagEmbedding import BGEM3FlagModel
 from pydantic import BaseModel
 from starlette.status import HTTP_504_GATEWAY_TIMEOUT
 
-os.environ["HF_HOME"] = "/tmp/cache"
+Path("/tmp/cache").mkdir(exist_ok=True)
+os.environ["HF_HOME"] = "/tmp/cache"  # does not quite work, need
 
 batch_size = 2 # gpu batch_size in order of your available vram
 max_request = 10 # max request for future improvements on api calls / gpu batches (for now is pretty basic)