deeppavlov / stand_docker_base

Base Dockerfile for the GPU based stand models

Installation and build

1. Install Docker according the instructions:

   https://docs.docker.com/install/linux/docker-ce/ubuntu/

2. Install nVIDIA Docker according the instructions:

   https://github.com/NVIDIA/nvidia-docker

3. Clone the repo and cd to project root:

   git clone https://github.com/deepmipt/stand_docker_base.git
   cd stand_docker_base
   

4. Build base stand Docker images for each CUDA versions:

   sudo docker build -t stand/docker_cuda:8.0 cuda/cuda_8.0/
   sudo docker build -t stand/docker_deeppavlov:cuda-8.0 deeppavlov/cuda_8.0/
   
   sudo docker build -t stand/docker_cuda:9.0 cuda/cuda_9.0/
   sudo docker build -t stand/docker_deeppavlov:cuda-9.0 deeppavlov/cuda_9.0/
   

5. Build DeepPavlov Jupyter server Docker images:

   sudo docker build -t stand/docker_cuda:9.0 cuda/cuda_9.0/
   sudo docker build -t stand/docker_deeppavlov:cuda-9.0 deeppavlov/cuda_9.0/
   sudo docker build -t stand/docker_deeppavlov_jupyter:cuda-9.0 deeppavlov/cuda_9.0_jupyter/
   

   Run DeepPavlov Jupyter server:

   sudo docker run -p 8888:8888 stand/docker_deeppavlov_jupyter:cuda-9.0
   

License

Apache 2.0 - licensed.