Update Dockerfile

Dockerfile

@@ -1,24 +1,22 @@
-# 1. Start from a community-vetted, GPU-ready image
-FROM runpod/pytorch:2.2.0-py3.10-cuda12.1.1-devel
+# 1. Start from the official vLLM OpenAI-compatible image.
+# This image is guaranteed to have the correct vLLM and CUDA dependencies.
+FROM vllm/vllm-openai:latest
 
-# 2. Set up the environment
-ENV DEBIAN_FRONTEND=noninteractive
-ENV HF_HOME="/data/huggingface"
-ENV UV_CACHE_DIR="/data/uv_cache"
+# 2. Override the default entrypoint of the base image so we can run our own code.
+ENTRYPOINT []
 
-# 3. The base image already has python, pip, etc. Install uv.
-RUN pip install uv
+# 3. Install the extra packages needed for our custom Gradio UI.
+# The base image already contains vllm, fastapi, pydantic, and torch.
+RUN pip3 install --no-cache-dir \
+    gradio==4.31.0 \
+    uvicorn
 
-# 4. Copy requirements and install them.
-COPY requirements.txt .
-RUN uv pip install --no-cache --system -r requirements.txt
-
-# 5. Copy the application code.
+# 4. Copy our custom application file.
 COPY ./app.py /app/app.py
 WORKDIR /app
 
-# 6. Expose the port.
+# 5. Expose the port our app will run on.
 EXPOSE 7860
 
-# 7. Run the application with uvicorn.
+# 6. Define the command to run our combined FastAPI/Gradio application using uvicorn.
 CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]