New Example Using YOLO

[heisenbuug]

Hey, as YOLO has been already added to the model repo I was thinking I can make a Face Mask Detector using YOLO. Shall I start working on it? It will be a good example for year 2020.

[kartikdutt18]

Hey there, That would be really interesting to add however training YOLO model takes a lot of time (one epoch takes approx 10 hours on 64x 64 image). You would have to first train the model in PyTorch (hopefully using GPU) and then use the mlpack-PyTorch-weight-translator. This is what used for transferring weights. If you have weights already you can test it out and I agree it would be a really good example.

[heisenbuug]

I will see if I can get weights from any other source, like any already implemented network in pytorch, then as suggested I can use mlpack-PyTorch-weight-translator and proceed. Will that be ok?

[kartikdutt18]

Sure, That makes sense.

[heisenbuug]

Hey @kartikdutt18 I found this repo
I think we can use these weights and dataset. What do you think?

[kartikdutt18]

Hey @heisenbuug, The PR for YOLOv3 hasn't be merged yet. Is there any dataset that uses YOLOv1?

[heisenbuug]

@kartikdutt18 oh, I will try finding some dataset that uses YOLOv1..I will update as soon as I find any...

[heisenbuug]

Hey @kartikdutt18 There is no dataset for YOLOv1, everyone is using YOLOv3 for face mask detection so we wont be getting any weights for v1 mostly. Any suggestions?

[mlpack-bot]

This issue has been automatically marked as stale because it has not had any recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions! 👍

[heisenbuug]

@kartikdutt18, any idea when will the PR for YOLOv3 will get merged?

[kumar-utkarsh0317]

Hey, I have started working on this, and I am having issues running the bash script (run.sh) of the mlpack-PyTorch weight converter repo
https://github.com/kartikdutt18/mlpack-PyTorch-Weight-Translator

I have raised an issue in that repo and hope to get some response soon so that I can proceed further