Please give an example to explain how to run this code.

[WisonHuang]

[TropComplique]

Hi! This is an unfinished implementation of this paper: https://arxiv.org/abs/1605.02019

How to run it:

1. preprocess_data.ipynb
2. get_windows.ipynb
3. train.ipynb - but the gradient for some reason explodes. I hope, next week I will fix this implementation.

[qiaochen]

Hi Антощенко,

Thanks for your effort to implement LDA2vec with pytorch. You might have noticed the gradient issue that results in the halt of total losses during training. I'm really curious about the cause of the issue, would you please also inform me of the details of the problem after fixing it? Thanks and good luck!

[TropComplique]

The cause of the issue was logarithm function. I fixed the problem by changing log(x) to log(x + epsilon) where epsilon is small.