Update os.environ["HF_HOME"]="/tmp/cache"

Dockerfile

@@ -22,5 +22,4 @@ RUN pip install --no-cache-dir --upgrade pip && \
 # CMD ["python", "app.py"]
 
 # CMD ["sh", "start-m3-server.sh"]
-
 CMD ["python", "m3_server.py"]

m3_server.py

@@ -1,3 +1,4 @@
+impor os
 import asyncio
 import time
 from concurrent.futures import ThreadPoolExecutor
@@ -10,6 +11,8 @@ from FlagEmbedding import BGEM3FlagModel
 from pydantic import BaseModel
 from starlette.status import HTTP_504_GATEWAY_TIMEOUT
 
+os.environ["HF_HOME"] = "/tmp/cache"
+
 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)
 max_length = 5000 # max context length for embeddings and passages in re-ranker