Add functionality to have preconfigured notebooks for each challenge

[nikochiko]

I'm submitting a

(Check one with "x" from given options)

• Bug report
• Feature request

Current behavior

Google Colab notebooks have become quite popular now. Many researchers (potential users) are turning to Google Colab for their projects due to its accessibility, mobility and ease of use.
In the context of EvalAI, a participant user has to go through the documentation of the host and explore around to get an idea of the challenge, find the training and validation datasets and set up the prerequisites. This costs a lot of time, not to mention it may sometimes be hard to find and implement all the instructions without errors.

Expected behavior

We can make the challenge host submit an optional IPython notebook with the dataset / prerequisites set up beforehand. This will give a headstart to the participant user as they can try out the challenge using their favorite ML library (TensorFlow, PyTorch, Caffe, or Swift for TensorFlow) right away without any hassle.

Other information

• It may also be useful to add a feature to link notebooks to the user's profiles (also enabling multiple notebooks for same challenge).
• Online notebooks are one of the biggest perks of Kaggle and I believe it makes a big difference. See: https://www.kaggle.com/docs/kernels
• It might be a somewhat big task to undertake, therefore the work can be segregated into backend, frontend and DevOps after a detailed spike analysis.