Create Dockerfile

Dockerfile

@@ -0,0 +1,16 @@
+FROM ghcr.io/ggml-org/llama.cpp:server
+
+RUN apt update && apt install wget -y && rm -rf /var/lib/apt/lists/*
+
+# Download the Q4_K_M model for faster CPU inference
+RUN wget "https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF/resolve/main/gemma-4-E2B-it-Q4_K_M.gguf" -O /model.gguf
+
+# Download the multimodal projector (required for Image & Audio)
+RUN wget "https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF/resolve/main/mmproj-gemma-4-E2B-it-Q4_K_M.gguf" -O /mmproj.gguf
+
+# -t 2 matches your 2vCPU resources exactly
+# --mmproj enables the vision/audio features
+CMD ["-m", "/model.gguf", \
+     "--mmproj", "/mmproj.gguf", \
+     "--port", "7860", "--host", "0.0.0.0", \
+     "-t", "2", "--mlock", "-b", "512"]