about VOC2007 dataset training

sovit-123 / fasterrcnn-pytorch-training-pipeline

PyTorch Faster R-CNN Object Detection on Custom Dataset

MIT License

223 stars 75 forks source link

about VOC2007 dataset training #120

Closed CharisWg closed 10 months ago

CharisWg commented 10 months ago

I do not understand what you mean TRAIN_DIR_IMAGES: '../input/voc_07_12/voc_xml_dataset/train/images' Are the XMLs in the Annotation files? or the text in the ./fasterrcnn-pytorch-training-pipeline-main/data/VOC2007/ImageSets/Segmentation/train.txt ???

sovit-123 commented 10 months ago

The code does not use text files to map the annotations for Pascal VOC or any other dataset. It uses XML annotation format for training. I think the training on custom dataset section will clarify your doubt.

CharisWg commented 10 months ago

I solved this, thank you so much for replying quickly.

sovit-123 commented 10 months ago

Welcome. Closing the issue for now.