transformers/docker/transformers-pytorch-gpu/Dockerfile
Adrien David-Sivelle 98fb718577
Docker GPU Images: Add NVIDIA/apex to the cuda images with pytorch (#7598)
- Use cuda:10.2 image instead of 10.1 (to address version mismatch
  warning with pytorch)
- Use devel version that is built on the runtime and includes headers
  and development tools (was otherwise failing to build apex)
2020-10-06 15:23:32 +02:00

31 lines
825 B
Docker

FROM nvidia/cuda:10.2-cudnn7-devel-ubuntu18.04
LABEL maintainer="Hugging Face"
LABEL repository="transformers"
RUN apt update && \
apt install -y bash \
build-essential \
git \
curl \
ca-certificates \
python3 \
python3-pip && \
rm -rf /var/lib/apt/lists
RUN python3 -m pip install --no-cache-dir --upgrade pip && \
python3 -m pip install --no-cache-dir \
mkl \
torch
RUN git clone https://github.com/NVIDIA/apex
RUN cd apex && \
python3 setup.py install && \
pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./
WORKDIR /workspace
COPY . transformers/
RUN cd transformers/ && \
python3 -m pip install --no-cache-dir .
CMD ["/bin/bash"]