search

astra-vision / ManiFest

Few-shot image translation method working on unstructured environments. ECCV 2022
Apache License 2.0
49 stars 4 forks source link

How long for training in a single GPU? #5

Closed ShenZheng2000 closed 1 year ago

ShenZheng2000 commented 1 year ago

Hi, authors!

Thanks for your excellent work. I would like to know how long it takes to train your model on a single GPU (e.g., 3090Ti).

Besides, I would like to know if you have tested the inference speed and FLOPs on high-resolution (e.g., 2k) images. How is your computational efficiency compared with the previous state-of-the-art?

fabvio commented 1 year ago

Thanks! Training takes around one day with the training settings specified in the code. We didn't test computational efficiency since it wasn't in the scope of the paper, but it would be useful if you make a test to report the results here for future reference.