search

sovit-123 / fasterrcnn-pytorch-training-pipeline

PyTorch Faster R-CNN Object Detection on Custom Dataset
MIT License
223 stars 75 forks source link

I need validation loss plot #32

Open MheadHero opened 1 year ago

sovit-123 commented 1 year ago

Will plan on adding it. Have it working on some other important things like aspect ratio resizing and mosaic similar to Ultralytics YOLOv5. Next, I will try to target this.

NammaThalle commented 1 year ago

is this ready? or can you give any work around for this for now?

sovit-123 commented 1 year ago

@NammaThalle I was trying to figure this out. But the only way available introduces some duplicate code in the validation loop and also leads to repeating the validation forward pass to calculate the mAP. So, this feature is on hold for now.

EmmaVanPuyenbr commented 1 year ago

I have the same question :)

sovit-123 commented 1 year ago

@EmmaVanPuyenbr I will try to figure it out but cannot promise anything.