Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

YOLOv5: State-of-the-art object detection at 140 FPS

blog.roboflow.ai

Post reply on HN
Print view View original

71–80 of 136 posts

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#71

Post by arctangent »

catalogia wrote: post #69
rcpt wrote: post #67

Earlier quoted context omitted.

I love that the response to them is "you can you up,no can no bb" Learned a new phrase today.

Can you explain it? I can't figure out what that means.

Apparently it is Chinese internet slang meaning:

"If you can do it, then you go and do it. If you can’t do it, then don’t criticise others."

via: http://www.chinesetimeschool.com/zh-cn/articles/chinese-inte...

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#72

Post by kbenson »

catalogia wrote: post #69
rcpt wrote: post #67

Earlier quoted context omitted.

I love that the response to them is "you can you up,no can no bb" Learned a new phrase today.

Can you explain it? I can't figure out what that means.

Just found these.[1][2] That is pretty awful, if it's from a dev.

Edit: Although as yeldarb explains in a comment here[3], it's probably a bit more complicated than that.

1: https://www.urbandictionary.com/define.php?term=you%20can%20...

2: https://www.quora.com/Whats-the-meaning-of-you-can-you-up-no...

3: https://news.ycombinator.com/item?id=23478983

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#73

Post by dang »

eeZah7Ux wrote: post #49
rocauc wrote: post #2

EfficientDet was open sourced March 18 [1], YOLOv4 came out April 23 [2], and now YOLOv5 is out only 48 days later. In our initial look, YOLOv5 is 180% faster, 88% smaller, similarly accurate, and easier to use (native to PyTorch rather thank Darknet) than YOLOv4. [1] https://venturebeat.com/2020/03/18/google-ai-open-sources-ef... [2] https://arxiv.org/abs/2004.10934

> open sourced This is not a verb.

It's behaving like one:

https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#75

Post by salty_biscuits »

nharada wrote: post #58

I welcome forward progress in the field, but something about this doesn't sit right with me. The authors have an unpublished/unreviewed set of results and they're already co-opting the YOLO name (without the original author) for it and all of this to promote a company? I guess this was inevitable when there's so much money in ML but it definitely feels against the spirit of the academic research community that they'r…

Well, very unlikely to get the original author. He doesn't do that kind of thing anymore

https://twitter.com/pjreddie/status/1230524770350817280?s=19

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#76

Post by kuzee »

bArray wrote: post #63

I'm just going to call this out as bullshit. This isn't YOLOv5. I doubt they even did a proper comparison between their model and YOLOv4. Someone asked it to not be called YOLOv5 and their response was just awful [1]. They also blew off a request to publish a blog/paper detailing the network [2]. I filed a ticket to get to the bottom of this with the creators of YOLOv4: https://github.com/AlexeyAB/darknet/issues/5920…

Looks like ultralytics, not roboflow, is the one that named this model v5. Different people/companies.

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#77

Post by jjcon »

ma2rten wrote: post #66

In February 2020, PJ Reddie noted he would discontinue research in computer vision. He actually stopped working on it because of ethical concerns. I'm inspired that he made this principled choice despite being quite successful in this field. https://syncedreview.com/2020/02/24/yolo-creator-says-he-sto...

In other words he stopped working on a project and needed an excuse to virtue signal.

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#78

Post by bArray »

kuzee wrote: post #76
bArray wrote: post #63

I'm just going to call this out as bullshit. This isn't YOLOv5. I doubt they even did a proper comparison between their model and YOLOv4. Someone asked it to not be called YOLOv5 and their response was just awful [1]. They also blew off a request to publish a blog/paper detailing the network [2]. I filed a ticket to get to the bottom of this with the creators of YOLOv4: https://github.com/AlexeyAB/darknet/issues/5920…

Looks like ultralytics, not roboflow, is the one that named this model v5. Different people/companies.

Yep, I updated my GitHub comment with respect to what @josephofiowa said. I made an assumptions when seeing the same PR images/language being used.

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#79

Post by bArray »

ebg13 wrote: post #40

It looks like this is YOLOv4 implemented on PyTorch, not actually a new YOLO?

YOLO is a neural network, Darknet is the framework. Without both YOLOv4 and "YOLOv5" on the same framework, it makes it near impossible to make any kind of meaningful comparison.

Re: YOLOv5: State-of-the-art object detection at 140 FPS

#80

Post by bArray »

kbenson wrote: post #72
catalogia wrote: post #69

Earlier quoted context omitted.

Can you explain it? I can't figure out what that means.

Just found these.[1][2] That is pretty awful, if it's from a dev. Edit: Although as yeldarb explains in a comment here[3], it's probably a bit more complicated than that. 1: https://www.urbandictionary.com/define.php?term=you%20can%20... 2: https://www.quora.com/Whats-the-meaning-of-you-can-you-up-no... 3: https://news.ycombinator.com/item?id=23478983

> Edit: Although as yeldarb explains in a comment here[3],

> it's probably a bit more complicated than that.

Legally speaking I'm not sure anything wrong was really done here.

Morally speaking, it seems quite unethical. AlexeyAB has really been carrying the torch of the Darknet framework and the YOLO neural network for quite some time (with pjreddie effectively handing it over to him).

AlexeyAB has been providing support on pjreddie's abandoned repository (e.g. [1]) and actively working on improvements in a fork [2]. If you look at the contributors graphs, he really has been keeping the project alive [3] (vs Darknet by pjreddie [4]).

Probably the worse part in my opinion is that they have also seemingly bypassed the open source nature of the project. This is quite damning.

[1] https://github.com/pjreddie/darknet/issues/1900

[2] https://github.com/AlexeyAB/darknet

[3] https://github.com/AlexeyAB/darknet/graphs/contributors

[4] https://github.com/pjreddie/darknet/graphs/contributors

Post reply on HN