Update Dockerfile

Dockerfile

@@ -1,34 +1,41 @@
 FROM python:3.11-slim
 
-# Install system dependencies for sentencepiece and other C++ builds
+# 1. Install system dependencies
+# We keep build-essential for any C++ extensions needed by transformers/accelerate
 USER root
 RUN apt-get update && apt-get install -y \
     build-essential \
+    git \
     && rm -rf /var/lib/apt/lists/*
 
-# Setup user
+# 2. Setup user
 RUN useradd -m -u 1000 user
 ENV HOME=/home/user \
     PATH=/home/user/.local/bin:$PATH \
     PYTHONUNBUFFERED=1 \
     PYTHONDONTWRITEBYTECODE=1 \
-    HF_HOME=/home/user/.cache/huggingface
+    HF_HOME=/home/user/.cache/huggingface \
+    # Ensure torch doesn't try to use too many CPU threads before the GPU kicks in
+    OMP_NUM_THREADS=4
 
 WORKDIR $HOME/app
 
-# Ensure cache directory exists
+# 3. Create cache directories
+# 7B models are large (~15GB), so proper cache management is vital
 RUN mkdir -p $HF_HOME && chown -R user:user $HOME
 
 USER user
 
-# Install Python requirements
+# 4. Install Python requirements
+# Make sure 'spaces' and 'accelerate' are in your requirements.txt
 COPY --chown=user requirements.txt .
 RUN pip install --no-cache-dir --upgrade pip && \
     pip install --no-cache-dir -r requirements.txt
 
-# Copy application
+# 5. Copy application
 COPY --chown=user . .
 
+# Hugging Face ZeroGPU requires the app to run on port 7860
 EXPOSE 7860
 
 CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]