CooLLaMACEO commited on
Commit
494bc63
·
verified ·
1 Parent(s): 3718e07

Update Dockerfile

Browse files
Files changed (1) hide show
  1. Dockerfile +7 -7
Dockerfile CHANGED
@@ -3,18 +3,18 @@ FROM python:3.11-slim
3
 
4
  WORKDIR /app
5
 
6
- # We still need wget for the model, but we can skip the heavy compilers!
7
  RUN apt-get update && apt-get install -y wget && rm -rf /var/lib/apt/lists/*
8
 
9
- # 1. Install pre-built llama-cpp-python (CPU version)
10
- # This bypasses the "Building wheel..." stage entirely
11
- RUN pip install llama-cpp-python \
12
- --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
13
 
14
- # 2. Install your other web tools
15
  RUN pip install --no-cache-dir fastapi uvicorn[standard] requests
16
 
17
- # 3. Setup models
18
  RUN mkdir -p ./models
19
  RUN wget -q -O ./models/gpt-oss-20b-Q3_K_M.gguf \
20
  https://huggingface.co/unsloth/gpt-oss-20b-GGUF/resolve/main/gpt-oss-20b-Q3_K_M.gguf
 
3
 
4
  WORKDIR /app
5
 
6
+ # Install wget (needed for the model)
7
  RUN apt-get update && apt-get install -y wget && rm -rf /var/lib/apt/lists/*
8
 
9
+ # 1. Install PREBUILT llama-cpp-python directly via URL
10
+ # This specific wheel is for Linux x86_64 and Python 3.11
11
+ RUN pip install --no-cache-dir \
12
+ https://github.com/abetlen/llama-cpp-python/releases/download/v0.3.16/llama_cpp_python-0.3.16-cp311-cp311-manylinux_2_17_x86_64.whl
13
 
14
+ # 2. Install web dependencies
15
  RUN pip install --no-cache-dir fastapi uvicorn[standard] requests
16
 
17
+ # 3. Setup models directory and download GPT-OSS-20B
18
  RUN mkdir -p ./models
19
  RUN wget -q -O ./models/gpt-oss-20b-Q3_K_M.gguf \
20
  https://huggingface.co/unsloth/gpt-oss-20b-GGUF/resolve/main/gpt-oss-20b-Q3_K_M.gguf