A Dockerfile consist of major machine learning libraries, for students and researchers.
- Ubuntu 16.04
- Python 3.6 on Miniconda3
- GPU accelerated (CUDA 8.0 + cuDNN 5)
- numpy, scipy, scikit-learn, ...
- Jupyter
- HDF5
- OpenCV 3
- ImageMagick
- ...
- TensorFlow
- Theano
- Keras
- PyTorch
- Chainer
- Torch7
- Caffe
- Caffe2
- Docker
- CUDA-enabled GPUs
- CUDA Toolkit
- nvidia-docker
Pull the docker image from DockerHub
$ docker pull uetchy/machinelearning
$ docker run -v $PWD:/app -p 80:8888 -it uetchy/machinelearning jupyter
$ open http://$(docker-machine ip default)
default
is a Docker Machine name. you would replace with whatever you want.
Docker Remote API users beware: Volume mounting is not available which is connected from Remote API
$ docker run -it uetchy/machinelearning python
$ docker run -it uetchy/machinelearning
$ cd /usr/src
$ ls