Update Dockerfile

Dockerfile

@@ -1,24 +1,26 @@
 
-# Use the official Hugging Face Docker image
 FROM huggingface/transformers-pytorch-gpu:latest
 
-# Install required packages
+# Install with backoff-retry for robustness
 RUN pip install --no-cache-dir \
     gradio \
     torch \
     transformers \
     sentencepiece \
-    accelerate
+    accelerate \
+    huggingface-hub \
+    datasets \
+    tenacity  # For retry logic
+
+# Configure cache
+ENV TRANSFORMERS_CACHE="/cache/models"
+ENV HF_DATASETS_CACHE="/cache/datasets"
+RUN mkdir -p /cache/models /cache/datasets
 
-# Create working directory
 WORKDIR /app
 COPY . /app
 
-# Download model (alternative: mount HF cache volume)
-# RUN python -c "from transformers import AutoModelForCausalLM; AutoModelForCausalLM.from_pretrained('meta-llama/Llama-2-7b-chat-hf')"
-
-# Expose port
-EXPOSE 7860
+# Pre-cache during build (optional)
+# RUN python cache_to_repo.py
 
-# Run the application
 CMD ["python", "app.py"]