LVKinyanjui commited on
Commit
8b66575
·
1 Parent(s): a07bd39

Removed the flash attention package altogether

Browse files
Files changed (1) hide show
  1. Dockerfile +5 -12
Dockerfile CHANGED
@@ -3,23 +3,16 @@ FROM python:${PYTHON_VERSION}-slim as base
3
 
4
  WORKDIR /app
5
 
6
- RUN apt update \
7
- && apt install git -y \
8
- && rm -rf /var/lib/apt/lists/*
 
9
 
10
- # Install problematic flash attention from source
11
- RUN pip install git+https://github.com/Dao-AILab/flash-attention.git
12
 
13
  COPY requirements.txt .
14
  RUN python -m pip install --no-cache-dir -r requirements.txt
15
 
16
- # To avert Permision error with transformer and hf models
17
- RUN mkdir ./.cache
18
-
19
- # # Install ollama
20
- # RUN curl -fsSL https://ollama.com/install.sh | sh
21
-
22
- # Copy the source code into the container.
23
  COPY . .
24
 
25
  # Expose the port that the application listens on.
 
3
 
4
  WORKDIR /app
5
 
6
+ # # Install problematic flash attention from source
7
+ # RUN apt update \
8
+ # && apt install git -y \
9
+ # && rm -rf /var/lib/apt/lists/*
10
 
11
+ # RUN pip install git+https://github.com/Dao-AILab/flash-attention.git
 
12
 
13
  COPY requirements.txt .
14
  RUN python -m pip install --no-cache-dir -r requirements.txt
15
 
 
 
 
 
 
 
 
16
  COPY . .
17
 
18
  # Expose the port that the application listens on.