search

Bohao-Lee / CME

64 stars 16 forks source link

the nan's problem has not been solved #12

Open a07913838438 opened 3 years ago

a07913838438 commented 3 years ago

Thanks for your great projects and Thanks for share the code of your great work! I do this

steps=-1,500,40000,60000
scales=0.1,10,.1,.1

to

steps=-1,500,40000,60000
scales=0.01,100,.1,.1

but The problem remains,I don't know how to do it to solve this problem.Can you give me some advice? 2021-08-24 10-39-11屏幕截图

Bohao-Lee commented 3 years ago

Thanks for your great projects and Thanks for share the code of your great work! I do this

steps=-1,500,40000,60000
scales=0.1,10,.1,.1

to

steps=-1,500,40000,60000
scales=0.01,100,.1,.1

but The problem remains,I don't know how to do it to solve this problem.Can you give me some advice? 2021-08-24 10-39-11屏幕截图

It seems that you have modified the batch size. Maybe you can change

 0.01, 100

to 

0.05, 20

or other combinations.

a07913838438 commented 3 years ago

Thanks for your great projects and Thanks for share the code of your great work! I do this

steps=-1,500,40000,60000
scales=0.1,10,.1,.1

to

steps=-1,500,40000,60000
scales=0.01,100,.1,.1

but The problem remains,I don't know how to do it to solve this problem.Can you give me some advice? 2021-08-24 10-39-11屏幕截图

It seems that you have modified the batch size. Maybe you can change

 0.01, 100

to

0.05, 20

or other combinations.

thank you very much,I have been solve this problem,because of the learning rate is very high.

pujianjian commented 3 years ago

Now, the loss value is not nan, but the proposal is 0. 1629858673(1)

Bohao-Lee commented 3 years ago

Now, the loss value is not nan, but the proposal is 0. 1629858673(1)

After training a while, proposal won't be 0. It only occurs at the beginning of training.