gathierry / fashionai-keypointsdetectionofapparel Goto Github PK
View Code? Open in Web Editor NEWFashionAI Key Points Detection using CPN model in Pytorch
License: Apache License 2.0
FashionAI Key Points Detection using CPN model in Pytorch
License: Apache License 2.0
Hello, thank you very much for the excellent work.
There is an alpha constant multiplied with the heatmap, and in the configure file the constant is 100, did you get better accuracy with the constant ? I read a few repos and there is no alpha constant, which make me a little confused.
@gathierry , FashionAI-KeyPintsDetectionofApparel got over on 5th of july. I came across this competition last week only. Though i already signed up, i am unable to access their dataset probably because the competition is already over. Could you share the dataset with me if you still have it ?
It was excellent to see your code and results, thank you for sharing the training code.
However, I couldn't find FashionAI dataset from their site, seems dataset is no more available. Could you please share trained model for inference. It will be very helpful for us to test it.
Best,
Abhishek
There are 15 version of stage 2. I want to ask which one is the best one? Do all of them work?
Hi Shiyu, great repo! I have two questions for you:
In config.py
, what's your rationale behind using 16
for self.hm_sigma = self.img_max_size / self.hm_stride / 16.
? How do you pick that number?
In stage2/keypoint_encoder.py
, why do you need to reduce the input image size when calculating the Gaussian key point? In addition, why do you have to subtract 1.0
in kpts[:,:2] = (kpts[:,:2] - 1.0) / stride
?
Thanks!
FashionAI dataset links are invalid. Is there anyone who can provide it? thank you very much!
我在coco2014数据集上训练人体骨骼点验证不出效果。不知何故?
This looks great work. Can you please add a license file so others can use it?
Thank you guys so much for this amazing repo, it's very inspiring to me.
Currently, I'm writing a project doing keypoints detection on traffic cone (which is similar to cloth but objects are much simpler). With only one class and 7 named keypoints on traffic cone keypoints, my trained network is about 300MB big and the inference speed is about 9 images per second on my 8GB RAM GTX1080. That means, the architecture is too deep here and the detection speed is kinda slow. The ideal detection speed would be 300 images per second under the same hardware.
Do you have any suggestions on how to modify the architecture to achieve that? Or maybe the Cascaded Pyramid Network is just too fancy for the traffic cone task?
Thank you so much!
Best Regards,
Sibo Zhu
could you please provide the check point file that has been generated for the code. And also could you provide the code to run the check point file to get the output on the image that we want to test( providing an unknown image and we need to get the keypoints of clothes on that image).
Can you give me the version of pytorch you use in this project ?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.