Mean Loss Val: nan

[Arjung27]

Hi Xingyul,

Thank you for providing your code here. I was trying to replicate the results from your paper and I am facing an issue doing that. When I run the code with the processed data you have provided I get my prediction scores as 'nan' (after 0.5 epoch only). As a result the loss value is 'nan' too. Please let me know what am I missing here?

Looking forward to your reply.

[s-ryosky]

Hi @Arjung27,
I found that there is NaN in the preprocessed flow data which are provided by authors.
You should skip the value in loading data.

[xingyul]

Thanks for bringing this up. I added one line to exclude this bad datapoint in 3a4b3b6.