Create Dockerfile
Browse files- Dockerfile +14 -0
Dockerfile
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
FROM ghcr.io/ggml-org/llama.cpp:full
|
| 2 |
+
|
| 3 |
+
WORKDIR /app
|
| 4 |
+
|
| 5 |
+
RUN apt-get update && apt-get install -y --no-install-recommends python3-pip python3-venv && rm -rf /var/lib/apt/lists/*
|
| 6 |
+
|
| 7 |
+
RUN python3 -m venv /opt/venv
|
| 8 |
+
ENV PATH=/opt/venv/bin:$PATH
|
| 9 |
+
|
| 10 |
+
RUN pip install --no-cache-dir -U huggingface_hub
|
| 11 |
+
|
| 12 |
+
RUN python3 -c 'from huggingface_hub import hf_hub_download; hf_hub_download(repo_id="yuxinlu1/gemma-4-12B-coder-fable5-composer2.5-v1-GGUF", filename="gemma4-coding-Q4_K_M.gguf", local_dir="/app")'
|
| 13 |
+
|
| 14 |
+
CMD ["--server","-m","gemma4-coding-Q4_K_M.gguf","--host","0.0.0.0","--port","7860","-t","3","--mmap","--threads-http","1","-c","16384","-n","4096"]
|