Update Dockerfile

Dockerfile

@@ -1,22 +1,22 @@
-FROM python:3.10-bullseye
+# Use the official pre-compiled image (CPU version)
+FROM ghcr.io/abetlen/llama-cpp-python:latest
 
 WORKDIR /app
 
-# Install only the essentials
+# 1. Install curl and move to the right folder
 RUN apt-get update && apt-get install -y curl && rm -rf /var/lib/apt/lists/*
 
-# Force PIP to use CPU-only binaries and avoid the build-stage entirely
-ENV PIP_PREFER_BINARY=1
-RUN pip install --no-cache-dir fastapi uvicorn
-RUN pip install llama-cpp-python \
-    --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
-
-# Download the model (MPT-7B Q2)
-# We do this AFTER the install so if the model fails, we don't have to reinstall tools
-RUN curl -L "https://huggingface.co/maddes8cht/mosaicml-mpt-7b-chat-gguf/resolve/main/mosaicml-mpt-7b-chat-Q2_K.gguf?download=true" -o mpt-7b-q2.gguf
+# 2. Download the model using the CLI (most stable)
+RUN pip install huggingface_hub
+RUN huggingface-cli download maddes8cht/mosaicml-mpt-7b-chat-gguf mosaicml-mpt-7b-chat-Q2_K.gguf --local-dir . --local-dir-use-symlinks False
+RUN mv mosaicml-mpt-7b-chat-Q2_K.gguf mpt-7b-q2.gguf
 
+# 3. Copy your app code
 COPY app.py .
 
+# Force the port to 7860 for Hugging Face
+ENV PORT=7860
 EXPOSE 7860
 
-CMD ["python", "app.py"]
+# Use the python inside the container to run your app
+CMD ["python3", "app.py"]