Alex Schultz

I'm hitting the same issue. I have my training data in "data/VOCdevkit/VOC2018/..." saving list to disk... Traceback (most recent call last):.. File "tools/prepare_dataset.py", line 118, in db.save_imglist(args.target, root=args.root_path) File "/home/aschu/development/apache-incubator/mxnet/incubator-mxnet/example/ssd/tools/../dataset/imdb.py",...

any update on this, I am also seeing very poor mAP after over 1000 epochs. I also tried using training data for validation and it was still very low.

@seveibar any chance you can look at this?