Video Analysis with Identity - Lowering memory footprint#2785
Merged
n-poulsen merged 4 commits intopytorch_dlcfrom Nov 14, 2024
Merged
Video Analysis with Identity - Lowering memory footprint#2785n-poulsen merged 4 commits intopytorch_dlcfrom
n-poulsen merged 4 commits intopytorch_dlcfrom
Conversation
2 tasks
maximpavliv
approved these changes
Nov 13, 2024
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR fixes a memory issue when running video analysis using a model with an identity head, as reported in #2771.
Improvements
PadOutputs: instead of needing to set"key": 0for outputs we don't want to pad (such asidentity_heatmap), check if each output is present in themax_individualsdict to see if padding is needed.PredictKeypointIdentities: Models with identity heads would output the identity heatmaps, which leads to memory issues when analyzing large videos. Akeep_id_maps=Falseparameter is added, only keep identity heatmaps in the output when explicitly requested.assign_identity: use the pose confidence to weight identity scores when assigning identity, as done in theTensorFlowimplementationBug fixes
make_pytorch_test_configwhen there are unique bodyparts for a project. The unique bodyparts need to be included inall_joint_namesandall_joints, otherwise video analysis fails.