This repository has been archived by the owner on Mar 21, 2024. It is now read-only.
Add subsampling transform and mean pooling #656
Merged
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
dccastro
changed the title
dccastro
changed the title
maxilse
approved these changes
Feb 21, 2022
harshita-s
reviewed
Feb 21, 2022
| @@ -204,7 +204,7 @@ def forward(self, instances: Tensor) -> Tuple[Tensor, Tensor]: # type: ignore | |||
| with set_grad_enabled(self.is_finetune): | |||
| instance_features = self.encoder(instances) # N X L x 1 x 1 | |||
| attentions, bag_features = self.aggregation_fn(instance_features) # K x N | K x L | |||
| bag_features = bag_features.view(-1, self.num_encoding * self.pool_out_dim) | |||
| bag_features = bag_features.view(1, -1) | |||
There was a problem hiding this comment.
why was this changed here?
There was a problem hiding this comment.
I believe this is a more robust way to reshape the outputs, without relying on num_encoding and pool_out_dim provided by the encoder and pooling components. In particular, MeanPoolingLayer ignores all arguments passed to it, so this operation would have failed if we expected a different shape here.
harshita-s
approved these changes
Feb 21, 2022
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Create a MONAI-style transform to allow subsampling tensors/arrays/lists to a given maximum length. This is to help enforce
max_bag_sizein a MIL setting after the full data has already been loaded (e.g. after caching).Piggybacking on this PR, support has also been added for mean pooling in the DeepMIL pipeline.