Ayoosh Kathuria

@anjackq Did the results of the training differ?

Scaling scales the image while keeping the dimensions constant. If you were to scale a image 2x, you will see a _zooming_ effect with some part of the image expelled...

This is a known issue. I'm busy at the moment. Won't be able to look it up until the end of this month. If you solve this, feel free to...

I'll try to solve your issue when I'll start working on the training code, which is soon. Sorry for the late reply. Been very busy this month.

I'm gonna write the train code very soon, and write another tutorial. Been really busy with this month with my college ending, moving stuff back home and looking for jobs.

How many images are we talking here? All the image addresses are loaded into a list called `imlist` before the detector is run. You think we might have an overflow,...

@skrish13 is right. @tomhayat Will implement a DataLoader class to load images in the coming week or so, then it wouldn't be a problem.

@munziliashali This has to do with the header file of the weights function. Very strange indeed since for the official YOLO v3 weights file supplied by Joseph Redmon, the header...

@munziliashali . Can you confirm the implementation can work on custom trained file, (like on your 17 class dataset?).

I guess the weights file is trained on COCO dataset, and not Pascal VOC? The dataset has an effect on the shape of feature map in the output code.