search

zsyOAOA / ResShift

ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting (NeurIPS@2023 Spotlight, TPAMI@2024)
Other
870 stars 47 forks source link

Training time #68

Open jasoaa11 opened 4 months ago

jasoaa11 commented 4 months ago

Dear Author Can I ask how long it took you to train the model on an NVIDIA Tesla V100 GPU?

zsyOAOA commented 4 months ago

I can't forget the exact time. For the conference version, about 1-2 days on 8 V100. For the journal version, about 4 days on 8 V100.

jasoaa11 commented 4 months ago

Thank you

XDeadlocked commented 3 months ago

I can't forget the exact time. For the conference version, about 1-2 days on 8 V100. For the journal version, about 4 days on 8 V100.

I can't forget the exact time. For the conference version, about 1-2 days on 8 V100. For the journal version, about 4 days on 8 V100.

Hello, is the batch_size and microbatch for the above training the same as the profile given in your repository (batch : [64,8], microbatch: 8)? If it is the same when using 8 V100 for training.

zsyOAOA commented 3 months ago

Yes. @XDeadlocked