James040 commited on
Commit
8ddfa49
·
verified ·
1 Parent(s): 9ce3351

Update Dockerfile

Browse files
Files changed (1) hide show
  1. Dockerfile +4 -20
Dockerfile CHANGED
@@ -10,26 +10,10 @@ ENV PATH="/opt/venv/bin:$PATH"
10
  RUN pip install -U pip huggingface_hub gradio requests
11
 
12
  # Download SmolLM2-1.7B-Instruct GGUF (Q4_K_M ~1GB, runs well on free CPU)
13
- RUN python3 -c '\
14
- from huggingface_hub import hf_hub_download; \
15
- hf_hub_download(\
16
- repo_id="bartowski/SmolLM2-1.7B-Instruct-GGUF", \
17
- filename="SmolLM2-1.7B-Instruct-Q4_K_M.gguf", \
18
- local_dir="/app"\
19
- )'
20
 
21
  COPY app.py /app/app.py
 
 
22
 
23
- # Start llama.cpp server in background, then launch Gradio
24
- CMD bash -c '\
25
- /llama-server \
26
- -m /app/SmolLM2-1.7B-Instruct-Q4_K_M.gguf \
27
- --host 0.0.0.0 \
28
- --port 8080 \
29
- -t 2 \
30
- -c 4096 \
31
- -n 2048 \
32
- --log-disable &
33
- sleep 10 && \
34
- python3 /app/app.py\
35
- '
 
10
  RUN pip install -U pip huggingface_hub gradio requests
11
 
12
  # Download SmolLM2-1.7B-Instruct GGUF (Q4_K_M ~1GB, runs well on free CPU)
13
+ RUN python3 -c 'from huggingface_hub import hf_hub_download; hf_hub_download(repo_id="bartowski/SmolLM2-1.7B-Instruct-GGUF", filename="SmolLM2-1.7B-Instruct-Q4_K_M.gguf", local_dir="/app")'
 
 
 
 
 
 
14
 
15
  COPY app.py /app/app.py
16
+ COPY start.sh /app/start.sh
17
+ RUN chmod +x /app/start.sh
18
 
19
+ CMD ["/app/start.sh"]