Jupyter Notebook + PyTorch + OpenCV (for Machine Learning projects) - No library installation required

https://hub.docker.com/r/diliplilaramani/ai_docker

Depends on base docker: jupyter/tensorflow-notebook

Libraries included - Python 3.7

1. Tensorflow = 2.0.0-beta1
2. Keras = 2.2
3. opencv-python==3.4.2.17
4. ipywidgets=7.4*
5. pandas=0.24*
6. numexpr=2.6*
7. matplotlib=3.0*
8. scipy=1.2*
9. seaborn=0.9*
10. scikit-learn=0.20*
11. scikit-image=0.14*
12. sympy=1.3*
13. cython=0.29*
14. patsy=0.5*
15. statsmodels=0.9*
16. cloudpickle=0.8*
17. dill=0.2*
18. dask=1.1.*
19. numba=0.42*
20. bokeh=1.0*
21. sqlalchemy=1.3*
22. hdf5=1.10*
23. h5py=2.9*
24. vincent=0.4.*
25. beautifulsoup4=4.7.*
26. protobuf=3.7.*
27. xlrd

Via command, docker-compose up (Recommended)

Step 1 - cd your_project

Step 2 - curl -O https://raw.githubusercontent.com/diliplilaramani/docker-ml-jupyter-notebook/master/docker-compose.yml

Step 3 - docker-compose up

Via dockerfile

1. Install docker
2. docker pull diliplilaramani/ai_docker:latest
3. docker run -p 8888:8888 diliplilaramani/ai_docker

Push our Dockerfile to DockerHub

1. create dockerfile
2. docker build -t ai_docker .
3. docker images
4. docker tag bb38976d03cf diliplilaramani/ai_docker:1.2
5. docker push diliplilaramani/ai_docker

Saving and loading images

1. docker save image_name > image_name.tar
2. docker load --input image_name.tar

Mount volume (folder data) into docker container

docker run -p 8888:8888 -v /home/dilip/Desktop/ML_Intel_analytics_vidhya_challenge:/tmp --rm --name ml_intel diliplilaramani/ai_docker:0.1

Errors

1. if Jupiter notebook running issues - https://stackoverflow.com/a/50920208/1012643

LINK - https://ropenscilabs.github.io/r-docker-tutorial/04-Dockerhub.html