Jiwon comments

Results 6 comments of


                                            Jiwon

multi gpu train bug

This problem was also mentioned in #316 and I’ve already fixed it locally. I’m planning to open a PR soon to share the fix.

Fix: stabilize test after distributed training completion

Hi @probicheaux , I just realized I hadn’t signed the CLA earlier — that’s done now ✅ Would you mind taking another look at this PR when you have a...

Fix MeanAverageRecall: compute mAR@K using top-K detections per image (COCO-compliant)

Thank you for the encouraging feedback and detailed notes, Gabriel 🙏 I’ve recently resumed work on this and will add the unit test you suggested (per-image K-filtering with synthetic data)...

Fix MeanAverageRecall: compute mAR@K using top-K detections per image (COCO-compliant)

I've added a simple unit test with synthetic data to validate the mAR@K calculation. **Test setup:** - **15 images** in total. - Each image has **≤ 5 bounding boxes/detections**. -...

Fix MeanAverageRecall: compute mAR@K using top-K detections per image (COCO-compliant)

Hi @SkalskiP , I just realized I hadn’t signed the CLA earlier — that’s done now ✅ Would you mind taking another look at this PR when you have a...

MeanAverageRecall does not follow COCO: mAR@K should use top-K detections per image, not globally

@aviralgarg05 Thanks for your interest. But I’ve already opened PR #1967 which resolves this issue, so it’s currently being reviewed.