collinprather / drsa Goto Github PK

View Code? Open in Web Editor NEW

20.0 20.0 4.0 1.37 MB

An adapted PyTorch implementation of a Deep Recurrent Survival Analysis model.

Home Page: https://collinprather.github.io/drsa/

License: Apache License 2.0

Jupyter Notebook 75.60% Python 23.68% Makefile 0.72%

deep-learning lstm-neural-networks machine-learning pytorch recurrent-neural-networks survival-analysis

drsa's People

Contributors

Stargazers

Watchers

Forkers

reanschwarzer1 kiminh yusheng0104 raymvp

drsa's Issues

Issue of installing drsa

which python version should be used for DRSA?
When I tried to install DRSA, I got an error on its dependency, torch==1.3.1.
I tried python3.6, 3.7, and 3.8, but none of them allows me to install torch1.3.1.
Which python version should I use?

Thanks,

Add reasonable defaults to `DRSA`

Currently, some of the initialization arguments (n_features, hidden_dim, n_layers), are all required arguments. It would be a more user-friendly API to add reasonable defaults to get users to training their model quicker.

How do you prepare the input dataset?

In your code, you have (batch_size, seq_len, n_features). I would like to have the data organized in a long format. Could you please give more details on data preparation? Thanks.

Enable `DRSA` to initialize with or without embeddings

Currently, DRSA must have valid embeddings in order to be initialized. For the sake of small, reproducible, explanatory examples, it should be possible to use it without embeddings.

Recommend Projects