when I try to train with pretrained model loaded, kernel dies

dreamquark-ai / tabnet

PyTorch implementation of TabNet paper : https://arxiv.org/pdf/1908.07442.pdf

https://dreamquark-ai.github.io/tabnet/

MIT License

2.55k stars 470 forks source link

when I try to train with pretrained model loaded, kernel dies #546

Closed nikhoj closed 18 hours ago

nikhoj commented 2 days ago

I am trying to train a model after pretraining. whenever it reaches to last epoch it shows the final log message that best epoch was X .... and then keep running. After sometime the kernel dies. It happens always.

Could you please let me know if this is any mistake from my end? I was using in AWS Sagemaker.

nikhoj commented 2 days ago

when drop_last is set to 'True', this problem does not arise.

Optimox commented 18 hours ago

Well that's weird, glad you found a solution to this problem.