YOLO: Real-Time Object Detection • Hacker News Board

Re: YOLO: Real-Time Object Detection

#31

Post by stared » Tue, Apr 02, 2019, 9:16 PM UTC

Having worked with YOLO, I really recommend this intro: https://blog.paperspace.com/how-to-implement-a-yolo-object-d.... And in general, YOLO is performant and at the same time, it has a simpler architecture than the Fast(er) R-CNN family.

And in general, due to its head, it is WAY more readable in PyTorch than in TensorFlow; to the point, I use it as an example in Keras vs PyTorch example https://deepsense.ai/keras-or-pytorch/ (was here at some point).

Re: YOLO: Real-Time Object Detection

#32

Post by joshvm » Tue, Apr 02, 2019, 9:17 PM UTC

dacox wrote: post #8

Great project, but pretty old now.

Yolov3 is about a year old and is still state of the art for all meaningful purposes. It's fast and works well. You might get "better" results with a Faster RCNN variant, but it's slow and the difference will likely be imperceptible. Using map50 as pjreddie points out, isn't a great metric for object detection.

Re: YOLO: Real-Time Object Detection

#33

Post by detaro » Tue, Apr 02, 2019, 9:19 PM UTC

dacox wrote: post #8

Great project, but pretty old now.

recommendations for similarly easy but better/more "modern" alternatives?

Re: YOLO: Real-Time Object Detection

#34

Post by joshvm » Tue, Apr 02, 2019, 9:22 PM UTC

syntaxing wrote: post #6

Suprised to see this here since YOLO has been out for a while now. Shameless plug, I wrote an article on how to use use transfer learning on your custom dataset with the pretrained weights [1]. One of the downside of YOLO is that it uses his own deep learning library darknet. I find that the Tensorflow port dark flow easier to use but it haven't seen a v3 port yet. [1] https://www.powu3.com/ml/yolo/

There is a pytorch port from Ultralytics (https://github.com/ultralytics/yolov3). Nobody seems to have figured out how to achieve the training performance of darknet though, which is entirely uncommented C. The source is all there, but the loss function changed between v2 and v3, and its not documented in the paper. I think it's been fixed in that pytorch port now though. The only frustrating thing is that every commit in the repo is called update...

Alternatively... you can train in darknet and then run inference in another framework of choice.

Also shameless plug: I wrote an annotation tool which is designed to output darknet formatted labels: https://github.com/jveitchmichaelis/deeplabel

Re: YOLO: Real-Time Object Detection

#35

Post by indutny » Tue, Apr 02, 2019, 9:26 PM UTC

amelius wrote: post #2

Can we also get the orientation of each detected object?

With some changes - yes. I did this in my experimental project: https://github.com/indutny/resistenz/blob/master/python/mode...

The idea is to add an extra 2 params to the output of each classifier cell. Then do L2 normalization on them ( https://github.com/indutny/resistenz/blob/master/python/mode... ) and treat them as a cosine/sine pair.

The loss in this case would be the Euclidean distance between the actual and predicted pairs, which is equal to "2 * (1 - cos(x-y))".

Re: YOLO: Real-Time Object Detection

#36

Post by joshvm » Tue, Apr 02, 2019, 9:28 PM UTC

detaro wrote: post #33

dacox wrote: post #8

Great project, but pretty old now.

recommendations for similarly easy but better/more "modern" alternatives?

See Faster-RCNN, R-FCN, SSD, etc

I've ignored mask RCNN becuase it's significantly more time consuming to label your data.

The main candidates are all found in Facebook’s Detectron package, but they didn't feel it necessary to document anything in any significant level of detail: https://github.com/facebookresearch/Detectron

Re: YOLO: Real-Time Object Detection

#37

Post by IshKebab » Tue, Apr 02, 2019, 9:33 PM UTC

daenz wrote: post #25

YOLO, no! https://i.imgur.com/R1RZ2N0.png Jokes aside, we need better temporal consistency, especially when we start arming AI. citizen -> citizen -> citizen -> armed insurgent

The problem there isn't temporal consistency (although I agree that often sucks), it is over-reliance on context. The invisible sheep problem: http://aiweirdness.com/post/171451900302/do-neural-nets-drea...

Re: YOLO: Real-Time Object Detection

#38

Post by natvert » Tue, Apr 02, 2019, 9:42 PM UTC

Interesting this is trending now. We have actually just recently released an improved version of YoloV3 (called G-Darknet) https://github.com/generalized-iou/g-darknet, using GIoU as a loss, which is described here: https://giou.stanford.edu

Also notable in G-Darknet are some tools useful for training (called darkboard), see https://github.com/generalized-iou/g-darknet/tree/master/dar...

Re: YOLO: Real-Time Object Detection

#39

Post by pjreddie » Tue, Apr 02, 2019, 9:43 PM UTC

I don’t really understand how this is any different from Overfeat or SSD...

Re: YOLO: Real-Time Object Detection

#40

Post by canada_dry » Tue, Apr 02, 2019, 11:07 PM UTC

airstrike wrote: post #22

syntaxing wrote: post #18

Earlier quoted context omitted.

You should read his site, his IDGAF attitude is pretty funny. The FAQ section is the most entertaining.

Nothing beats this resume https://pjreddie.com/static/Redmon%20Resume.pdf

Like executives that shun computers as a symbol of their power and status... having a resume that emphasizes my-little-pony suggests his dance card is full (i.e. he has his pick of top job opportunities).