2D FCMAE #71
Merged
2D FCMAE #71
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
new black version has different rules
* inital commit adding the normalization. * adding dataset_statistics to each fov to facilitate the configurable augmentations * fix indentation * ruff * test preprocessing * remove redundant field * cleanup --------- Co-authored-by: Ziwen Liu <[email protected]>
Closed
mattersoflight
added this to the v0.1.0 release - related to mantis paper, robust VS preprint, and DL@MBL exercise milestone
edyoshikun
reviewed
Jun 5, 2024
| self._log_samples("val_samples", self.validation_step_outputs) | ||
| self.validation_step_outputs = [] | ||
| # average within each dataloader |
There was a problem hiding this comment.
Here I know that this is the end of the validation meaning that it won't backprop and therefore maybe no need to detach the tensor before doing any logging. Is this the common practice? I know that for the train_step and validation_step is more relevant and important If we don't detach here it doesnt affect it any way? Just curious..
There was a problem hiding this comment.
To log the loss value (through lightning logger), the detaching is automatic. We only need to care about it when logging manually (images).
edyoshikun
approved these changes
Jun 11, 2024
There was a problem hiding this comment.
I tested doing the following without issues:
- 2D predictions using a549 2D data
- 3D predictions using mantis A549 data
I don't expect it to break anything related to the dataloaders since I was using the hcs.py. I like the prepare_data_per_node flags for caching. I will just need to update my scripts to add this flag from now on.
Thank you so much for this Ziwen!
edyoshikun
added a commit
that referenced
this pull request
Jun 12, 2024
* refactor data loading into its own module * update type annotations * move the logging module out * move old logging into utils * rename tests to match module name * bump torch * draft fcmae encoder * add stem to the encoder * wip: masked stem layernorm * wip: patchify masked features for linear * use mlp from timm * hack: POC training script for FCMAE * fix mask for fitting * remove training script * default architecture * fine-tuning options * fix cli for finetuning * draft combined data module * fix import * manual validation loss reduction * update linting new black version has different rules * update development guide * update type hints * bump iohub * draft ctmc v1 dataset * update tests * move test_data * remove path conversion * configurable normalizations (#68) * inital commit adding the normalization. * adding dataset_statistics to each fov to facilitate the configurable augmentations * fix indentation * ruff * test preprocessing * remove redundant field * cleanup --------- Co-authored-by: Ziwen Liu <[email protected]> * fix ctmc dataloading * add example ctmc v1 loading script * changing the normalization and augmentations default from None to empty list. * invert intensity transform * concatenated data module * subsample videos * livecell dataset * all sample fields are optional * fix multi-dataloader validation * lint * fixing preprocessing for varying array shapes (i.e aics dataset) * update loading scripts * fix CombineMode * always use untrainable head for FCMAE * move log values to GPU before syncing Lightning-AI/pytorch-lightning#18803 * custom head * ddp caching fixes * fix caching when using combined loader * compose normalizations for predict and test stages * black * fix normalization in example config * fix normalization in example config * prefetch more in validation * fix collate when multi-sample transform is not used * ddp caching fixes * fix caching when using combined loader * typing fixes * fix test dataset * fix invert transform * add ddp prepare flag for combined data module * remove redundant operations * filter empty detections * pass trainer to underlying data modules in concatenated * hack: add test dataloader for LiveCell dataset * test datasets for livecell and ctmc * fix merge error * fix merge error * fix mAP default for over 100 detections * bump torchmetric * fix combined loader training for virtual staining task * fix non-combined data loader training * add fcmae to graph script * fix type hint * format * add back convolutiuon option for fcmae head --------- Co-authored-by: Eduardo Hirata-Miyasaki <[email protected]>
edyoshikun
added a commit
that referenced
this pull request
Jun 12, 2024
* refactor data loading into its own module * update type annotations * move the logging module out * move old logging into utils * rename tests to match module name * bump torch * draft fcmae encoder * add stem to the encoder * wip: masked stem layernorm * wip: patchify masked features for linear * use mlp from timm * hack: POC training script for FCMAE * fix mask for fitting * remove training script * default architecture * fine-tuning options * fix cli for finetuning * draft combined data module * fix import * manual validation loss reduction * update linting new black version has different rules * update development guide * update type hints * bump iohub * draft ctmc v1 dataset * update tests * move test_data * remove path conversion * configurable normalizations (#68) * inital commit adding the normalization. * adding dataset_statistics to each fov to facilitate the configurable augmentations * fix indentation * ruff * test preprocessing * remove redundant field * cleanup --------- Co-authored-by: Ziwen Liu <[email protected]> * fix ctmc dataloading * add example ctmc v1 loading script * changing the normalization and augmentations default from None to empty list. * invert intensity transform * concatenated data module * subsample videos * livecell dataset * all sample fields are optional * fix multi-dataloader validation * lint * fixing preprocessing for varying array shapes (i.e aics dataset) * update loading scripts * fix CombineMode * always use untrainable head for FCMAE * move log values to GPU before syncing Lightning-AI/pytorch-lightning#18803 * custom head * ddp caching fixes * fix caching when using combined loader * compose normalizations for predict and test stages * black * fix normalization in example config * fix normalization in example config * prefetch more in validation * fix collate when multi-sample transform is not used * ddp caching fixes * fix caching when using combined loader * typing fixes * fix test dataset * fix invert transform * add ddp prepare flag for combined data module * remove redundant operations * filter empty detections * pass trainer to underlying data modules in concatenated * hack: add test dataloader for LiveCell dataset * test datasets for livecell and ctmc * fix merge error * fix merge error * fix mAP default for over 100 detections * bump torchmetric * fix combined loader training for virtual staining task * fix non-combined data loader training * add fcmae to graph script * fix type hint * format * add back convolutiuon option for fcmae head --------- Co-authored-by: Eduardo Hirata-Miyasaki <[email protected]>
edyoshikun
added a commit
that referenced
this pull request
Jun 12, 2024
* refactor data loading into its own module * update type annotations * move the logging module out * move old logging into utils * rename tests to match module name * bump torch * draft fcmae encoder * add stem to the encoder * wip: masked stem layernorm * wip: patchify masked features for linear * use mlp from timm * hack: POC training script for FCMAE * fix mask for fitting * remove training script * default architecture * fine-tuning options * fix cli for finetuning * draft combined data module * fix import * manual validation loss reduction * update linting new black version has different rules * update development guide * update type hints * bump iohub * draft ctmc v1 dataset * update tests * move test_data * remove path conversion * configurable normalizations (#68) * inital commit adding the normalization. * adding dataset_statistics to each fov to facilitate the configurable augmentations * fix indentation * ruff * test preprocessing * remove redundant field * cleanup --------- Co-authored-by: Ziwen Liu <[email protected]> * fix ctmc dataloading * add example ctmc v1 loading script * changing the normalization and augmentations default from None to empty list. * invert intensity transform * concatenated data module * subsample videos * livecell dataset * all sample fields are optional * fix multi-dataloader validation * lint * fixing preprocessing for varying array shapes (i.e aics dataset) * update loading scripts * fix CombineMode * always use untrainable head for FCMAE * move log values to GPU before syncing Lightning-AI/pytorch-lightning#18803 * custom head * ddp caching fixes * fix caching when using combined loader * compose normalizations for predict and test stages * black * fix normalization in example config * fix normalization in example config * prefetch more in validation * fix collate when multi-sample transform is not used * ddp caching fixes * fix caching when using combined loader * typing fixes * fix test dataset * fix invert transform * add ddp prepare flag for combined data module * remove redundant operations * filter empty detections * pass trainer to underlying data modules in concatenated * hack: add test dataloader for LiveCell dataset * test datasets for livecell and ctmc * fix merge error * fix merge error * fix mAP default for over 100 detections * bump torchmetric * fix combined loader training for virtual staining task * fix non-combined data loader training * add fcmae to graph script * fix type hint * format * add back convolutiuon option for fcmae head --------- Co-authored-by: Eduardo Hirata-Miyasaki <[email protected]>
edyoshikun
added a commit
that referenced
this pull request
Jun 18, 2024
* refactor data loading into its own module * update type annotations * move the logging module out * move old logging into utils * rename tests to match module name * bump torch * draft fcmae encoder * add stem to the encoder * wip: masked stem layernorm * wip: patchify masked features for linear * use mlp from timm * hack: POC training script for FCMAE * fix mask for fitting * remove training script * default architecture * fine-tuning options * fix cli for finetuning * draft combined data module * fix import * manual validation loss reduction * update linting new black version has different rules * update development guide * update type hints * bump iohub * draft ctmc v1 dataset * update tests * move test_data * remove path conversion * configurable normalizations (#68) * inital commit adding the normalization. * adding dataset_statistics to each fov to facilitate the configurable augmentations * fix indentation * ruff * test preprocessing * remove redundant field * cleanup --------- Co-authored-by: Ziwen Liu <[email protected]> * fix ctmc dataloading * add example ctmc v1 loading script * changing the normalization and augmentations default from None to empty list. * invert intensity transform * concatenated data module * subsample videos * livecell dataset * all sample fields are optional * fix multi-dataloader validation * lint * fixing preprocessing for varying array shapes (i.e aics dataset) * update loading scripts * fix CombineMode * always use untrainable head for FCMAE * move log values to GPU before syncing Lightning-AI/pytorch-lightning#18803 * custom head * ddp caching fixes * fix caching when using combined loader * compose normalizations for predict and test stages * black * fix normalization in example config * fix normalization in example config * prefetch more in validation * fix collate when multi-sample transform is not used * ddp caching fixes * fix caching when using combined loader * typing fixes * fix test dataset * fix invert transform * add ddp prepare flag for combined data module * remove redundant operations * filter empty detections * pass trainer to underlying data modules in concatenated * hack: add test dataloader for LiveCell dataset * test datasets for livecell and ctmc * fix merge error * fix merge error * fix mAP default for over 100 detections * bump torchmetric * fix combined loader training for virtual staining task * fix non-combined data loader training * add fcmae to graph script * fix type hint * format * add back convolutiuon option for fcmae head --------- Co-authored-by: Eduardo Hirata-Miyasaki <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add option to not perform convolution in the head of the FCMAE models.
Fixes for multi-dataset trainer.
Fix default when computing metrics on images with very high instance counts.