search

yang-jin-hai / SAIN

[AAAI-2023] Self-Asymmetric Invertible Network for Compression-Aware Image Rescaling
Apache License 2.0
33 stars 2 forks source link

can you share the training log? #4

Open Feynman1999 opened 1 year ago

Feynman1999 commented 1 year ago

Hello, may I kindly ask if you can share the training logs? This would be helpful for me to replicate your work. Thank you! Additionally, have you tried adjusting the weight ratios of different losses, and does it have a significant impact on the results?

yang-jin-hai commented 1 year ago

You can refer to #1 first.

Feynman1999 commented 1 year ago

You can refer to #1 first.

yes, i have refer it, at the same setting, I am currently training and I don't know the final result yet. But the phenomenon I have observed is that sometimes the loss suddenly increases by about twice, and then the validation set results decrease, but later converge again. Have you ever encountered this phenomenon?

yang-jin-hai commented 1 year ago

The validation results would generally increase steadily and may sometimes drop slightly. But the losses may be unstable.