deploy

Dockerfile

@@ -4,11 +4,15 @@ FROM python:3.10-slim
 # Set the working directory inside the container
 WORKDIR /app
 
+# Set environment variable for Hugging Face cache directory
+ENV TRANSFORMERS_CACHE="/app/cache"
+
 # Copy the requirements file
 COPY requirements.txt .
 
-# Install dependencies
-RUN pip install --no-cache-dir -r requirements.txt
+# Install dependencies and update huggingface_hub
+RUN pip install --no-cache-dir -r requirements.txt \
+    && pip install --upgrade huggingface_hub
 
 # Copy the entire app to the container
 COPY . .
@@ -17,4 +21,4 @@ COPY . .
 EXPOSE 7860
 
 # Run FastAPI with Uvicorn
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

main.py

@@ -3,6 +3,8 @@ from pydantic import BaseModel
 from sentence_transformers import SentenceTransformer
 import numpy as np
 
+import os
+os.environ["TRANSFORMERS_CACHE"] = "/app/cache"
 # Initialize FastAPI app
 app = FastAPI()
 

requirements.txt

@@ -6,4 +6,5 @@ scikit-learn==1.2.2
 sentence-transformers==2.2.2
 numpy==1.23.5
 pydantic==1.10.8
-torch==2.0.1  # Required for sentence-transformers
+torch==2.0.1  # Required for sentence-transformers
+huggingface_hub>=0.16.4