Update Dockerfile

Dockerfile

@@ -1,36 +1,20 @@
-FROM python:3.10-slim
-
+FROM ghcr.io/ggml-org/llama.cpp:server
 USER root
-WORKDIR /app
-
-# 1. Install system basics
-RUN apt-get update && apt-get install -y curl && rm -rf /var/lib/apt/lists/*
-
-# 2. Install AI Stack
-# We use --extra-index-url so pip can find both Torch (CPU) and standard packages (FastAPI)
-RUN pip3 install --no-cache-dir --upgrade pip && \
-    pip3 install --no-cache-dir \
-    uvicorn \
-    fastapi \
-    transformers \
-    accelerate \
-    safetensors \
-    torch --extra-index-url https://download.pytorch.org/whl/cpu
-
-# 3. Create model directory and download GPT-2 XL Safetensors
-RUN mkdir -p /models/gpt2-xl && \
-    curl -L https://huggingface.co/openai-community/gpt2-xl/resolve/main/model.safetensors -o /models/gpt2-xl/model.safetensors && \
-    curl -L https://huggingface.co/openai-community/gpt2-xl/resolve/main/config.json -o /models/gpt2-xl/config.json && \
-    curl -L https://huggingface.co/openai-community/gpt2-xl/resolve/main/vocab.json -o /models/gpt2-xl/vocab.json && \
-    curl -L https://huggingface.co/openai-community/gpt2-xl/resolve/main/merges.txt -o /models/gpt2-xl/merges.txt && \
-    curl -L https://huggingface.co/openai-community/gpt2-xl/resolve/main/tokenizer.json -o /models/gpt2-xl/tokenizer.json && \
-    curl -L https://huggingface.co/openai-community/gpt2-xl/resolve/main/generation_config.json -o /models/gpt2-xl/generation_config.json
-
-# 4. Copy your server code
-COPY app.py .
-RUN chown -R 1000:1000 /app /models
+RUN apt-get update && apt-get install -y curl
+RUN mkdir -p /models && \
+    curl -L https://huggingface.co/unsloth/gemma-3n-E2B-it-GGUF/resolve/main/gemma-3n-E2B-it-Q8_0.gguf -o /models/model.gguf && \
+    chown -R 1000:1000 /models
 USER 1000
-
-EXPOSE 8080
-
-CMD ["python3", "app.py"]
+ENV LLAMA_ARG_MODEL=/models/model.gguf
+ENV LLAMA_ARG_HOST=0.0.0.0
+ENV LLAMA_ARG_PORT=7860
+ENV LLAMA_ARG_THREADS=8
+ENV LLAMA_ARG_BATCH_SIZE=2048
+ENV LLAMA_ARG_UBATCH_SIZE=512
+ENV LLAMA_ARG_CTX_SIZE=8196 
+ENV LLAMA_ARG_FLASH_ATTN=true
+ENV LLAMA_ARG_NO_MMAP=false 
+ENV LLAMA_ARG_MLOCK=true 
+HEALTHCHECK --interval=30s --timeout=15s --start-period=10s --retries=3 \
+    CMD curl -f http://localhost:7860/health || exit 1
+ENTRYPOINT ["/app/llama-server"]