Loss is NaN issue - Githubissues

shenghanlin / SeismicFoundationModel

Seismic Foundation Model

67 stars 16 forks source link

Loss is NaN issue #6

Closed TaniMasanori closed 7 months ago

TaniMasanori commented 8 months ago

Hi, When I try to do Finetune using Network after Pretrain, Loss is NaN before Epoch 0 starts. Please let me know if there is a solution for this kind of problem.

Thank you for sharing your great work!

shenghanlin commented 7 months ago

In this case, you may need to check whether there are problems with your data, such as non-normalization, and problems with NaN values.

TaniMasanori commented 7 months ago

Thank you! Now, I can run the training.