Create Dockerfile

Dockerfile

@@ -0,0 +1,45 @@
+# Use a base Ubuntu image
+FROM ubuntu:22.04
+
+# Set environment variables to prevent interactive prompts during apt operations
+ENV DEBIAN_FRONTEND=noninteractive
+
+# Install system dependencies: curl for Ollama, wget for model download, git (optional, but good for debugging), python3 and pip for the application
+RUN apt update && apt install -y curl wget git python3 python3-pip
+
+# Install Ollama using its official installation script
+# This script automatically detects the system architecture and installs the correct binary.
+RUN curl -fsSL https://ollama.com/install.sh | sh
+
+# Set the working directory inside the container
+WORKDIR /app
+
+# Define environment variables for the model repository and filename
+# This makes it easy to change the model later without editing the RUN command directly.
+ENV MODEL_REPO="unsloth/gemma-3-4b-it-qat-GGUF"
+ENV MODEL_FILENAME="gemma-3-4b-it-qat.Q4_K_M.gguf"
+
+# Download the specific GGUF model file directly from Hugging Face Hub.
+# We use 'wget -O' to save the file with the desired filename in the current directory.
+# The 'resolve/main' path ensures we get the raw file content.
+RUN wget -O ${MODEL_FILENAME} https://huggingface.co/${MODEL_REPO}/resolve/main/${MODEL_FILENAME}
+
+# Copy the Modelfile, application script, Python requirements, and the startup script into the container
+COPY Modelfile .
+COPY app.py .
+COPY requirements.txt .
+COPY run.sh .
+
+# Install Python dependencies required by your Gradio application
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Make the startup script executable
+RUN chmod +x run.sh
+
+# Expose the port that your Gradio application will listen on.
+# Hugging Face Spaces typically use port 7860 for Gradio apps.
+EXPOSE 7860
+
+# Set the entrypoint for the container to execute our startup script.
+# This script will start Ollama and then your application.
+CMD ["./run.sh"]