Is it correct that Nan appears in the loss?

[cjfghddl0206]

I ran the code as written on github. However, after a certain point, the loss is all Nan. I think it's a loss of the dataset, so I recreated the pkl file with create_data.py, but Nan comes out as it is. Is it correct to run the training to the end even if Nan comes out?

[DezeZhao]

maybe you need to change the batch size, I met this situation as well.

[cjfghddl0206]

배치 크기를 변경해야 할 수도 있습니다. 이 상황도 만났습니다.

Did you change the batch size from how many to how many?

[DezeZhao]

8 to 4

[cjfghddl0206]

8 to 4

Where should I edit it? I can't see the batch size 8 part.

[zhaowenZhou]

8 to 4

Where should I edit it? I can't see the batch size 8 part.

samples_per_gpu