Adding support for pyramids #129
Conversation
edyoshikun
force-pushed
the
support_pyramid
branch
from
308392c to
8a137d4
Compare
edyoshikun
force-pushed
the
support_pyramid
branch
from
eb83d8c to
8a137d4
Compare
|
@edyoshikun thanks - this implementation seems backward compatible with previous configs for training and prediction for virtual staining and contrastive learning tasks. Can you check with the configs we are sharing with the preprint if they need any change (e.g., explicitly setting the scale = 0)? |
|
@mattersoflight we don't need to change the previous configs because by default the Dataloader will be set to "0" the highest resolution if the parameter is not provided in the config. For the preprint, I thought we were pointing them to |
|
I've tested the basic @edyoshikun I'm happy to leave this as a branch while we test this out end to end during the course. |
| @@ -104,6 +104,8 @@ class SlidingWindowDataset(Dataset): | |||
| :param ChannelMap channels: source and target channel names, | |||
| e.g. ``{'source': 'Phase', 'target': ['Nuclei', 'Membrane']}`` | |||
| :param int z_window_size: Z window size of the 2.5D U-Net, 1 for 2D | |||
| :param str pyramid_resolution: pyramid level. | |||
| defaults to 0 (full resolution) | |||
There was a problem hiding this comment.
| defaults to 0 (full resolution) | |
| defaults to "0" (full resolution) |
|
The code changes look good, but we should test that training actually works before merging. |
This PR adds the support for pyramids to address #125. This modifies the
HCSDataloader()