You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Using an RTX3090 Ti 24GB, train our-s, n_sequence=12 on the DAVIS dataset, but looking at the loss, we find that the Loss has been fluctuating wildly.
I'm in the same boat, I can't seem to converge when I train on the DVD dataset, have you solved this problem yet? I've tried adjusting the learning rate several times and it doesn't seem to work.
Using an RTX3090 Ti 24GB, train our-s, n_sequence=12 on the DAVIS dataset, but looking at the loss, we find that the Loss has been fluctuating wildly.
![image](https://private-user-images.githubusercontent.com/83681909/332000996-0fb72126-9a9b-4b5f-8898-68dcee314843.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.azfR_hi86fqQ9dFLAYZy0nlJJJNUdl4w9kP4ytO7XJo)
The text was updated successfully, but these errors were encountered: