Training results: result_4.3.0-nobackground - #27
Closed
MateoLostanlen wants to merge 6 commits into
Closed
Conversation
Both fetches pointed at yolo_train_val_corrected / yolo_test_corrected at v4.0.0-corrected. Those are frozen snapshots added by pyro-dataset commit 76eb464 and tracked outside its pipeline: their content hash is identical at v4.0.0-corrected, v4.1.0, v4.2.0 and v4.3.0. Bumping the tag alone would have retrained on the same 32201 files and reported it as a v4.3.0 model. Switches to the pipeline outputs, which do carry the annotator imports — roughly 18169 train+val images against 16100 in the snapshot, and the test set grows from 183 to 190 sequences. The trade is the frame-level annotation corrections from vision-rd/data-quality, which only ever existed for the v4.0.0 frames. Reproducing that pass on v4.3.0 would be the way to get both. The result-PR footer named the old tag too; it would have misreported which dataset the numbers came from.
Add --exclude-background flag to build.py that drops background images (empty label files) from the train split, and enable it in the train_yolo_best DVC stage. Removes 1642 background images from train (13039 remaining); val split kept intact for comparable evaluation.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Evaluation results on test set
Sequential evaluation on test (optimized params: nb_frames=6, conf=0.35)
TP=148 FN=7 FP=33 TN=109
Engine parameter optimization (top 10 on val)
(vs last committed results on main)
Branch:
result_4.3.0-nobackground| Test dataset: pyronear/pyro-dataset @ v4.3.0