search

mihirp1998 / Diffusion-TTA

Diffusion-TTA improves pre-trained discriminative models such as image classifiers or segmentors using pre-trained generative models.
https://diffusion-tta.github.io
Other
50 stars 4 forks source link

Questions about the implementation for the segmentation task #6

Open JdbermeoUZH opened 7 months ago

JdbermeoUZH commented 7 months ago

Hi,

Thank you very much for this work! I wanted to ask you a couple of quick questions about the adaptation for the segmentation task:

  1. Did you use SGD or ADAM for it?
  2. Was the TTA online or single sample for this task?

Again, thank you very much and congrats on the wonderful work

mihirp1998 commented 7 months ago

Hi,

The TTA was single sample and we used AdamW optimizer with learning rate of 8e-05.

JdbermeoUZH commented 7 months ago

Thanks!