subhc / unsup-parts Goto Github PK
View Code? Open in Web Editor NEWUnsupervised Part Discovery from Contrastive Reconstruction (NeurIPS 2021)
Home Page: https://www.robots.ox.ac.uk/~vgg/research/unsup-parts/
Unsupervised Part Discovery from Contrastive Reconstruction (NeurIPS 2021)
Home Page: https://www.robots.ox.ac.uk/~vgg/research/unsup-parts/
Hi, thanks for the great and very inspirational work. It helps me learn a lot from your def. of parts&object, motivation of your several different losses and ablation study!
But I am still a little confused about the visual consistency loss(in Sec.3.2), where you mentioned that the Gaussian model was used. However, I didn't find the Gaussian model in equation(4). Where do you utilize the Gaussian model? Could you generously send me(email: [email protected]) the code about how you implement the visual consistency loss in Sec.3.2? I really appreciate your work and help, thanks a lot!
the model seemed unable to converge
Hello,
Thank you for your interesting work!
We are currently conducting extensive research based on your contributions. However, we are facing some challenges with the evaluation process with the PASCAL-Parts dataset.
Would it be possible for you to share the evaluation code used for the PASCAL-Parts dataset? Access to this code would be incredibly beneficial for our research.
Thank you for considering our request.
Thanks for your great research, I think that this is a very novel work and it can help me improve my research. Therefore, I want to apply the code for this project, this is my google mail: [email protected]. Thank you very much!
Hi,
The models are missing.
python3 train.py
FileNotFoundError: [Errno 2] No such file or directory: 'models'
Hi @subhc ,
Thank you for your great work!
I wonder how to get the key point predicted labels or part labels from your code (especially, the file CUB_eval.ipynb)
Could you please help me with that?
Thank you very much
Tin
Thank you for your interesting work and for sharing your code in the future. Is that possible to ask when you will upload the source code? Thank you very much.
Hello,
In the file models/feature_extraction.py line 46, you are importing vgg16 from utils.deepcluster_vgg16, but I am unable to find it anywhere. It is not under the utils directory in the root folder. I also searched the repo, but I could not find any mentions of deepcluster_vgg16. Can you please help me with this? Many thanks!
Hello, thank you for your inspiring work! I tried to re-run the code on all of the datasets, but the results were not as promising as those repored by paper.
For CUB
and DeepFashion
, I did not modify any of the codes, but the metrics on CUB
were poor.
And for Pascal-Part
, I modified the training hyper-params according to your supp file. And also for the fairness of evaluation, I trained a foreground segmentator (a DeepLabV2-ResNet50-2branch
) and used the predicted mask at evaluation. I conducted training upon Car
, Cat
and Horse
. The results on Horse
were OK, but those on Cat
and Car
were really not good.
Could you please provide analysis upon why the re-run results on CUB
were not good? Also, for Pascal-Part
, are there any training details that is left out in papers, so that I did not reproduce the results? (BTW, could you provide supervised mask of PascalPart
?)
How are the 15 parts from the ground truth keypoints mapped to just 4 parts to valuate the model?
Thanks for the great work and your effort in releasing the code! Is there an expected date that your code will be released? Thanks in advance.
I am trying to experiment with batch size of 2 but seems like there is an index error in loss.py
Hi, I am trying to train the network on a custom dataset of shoes. After training overnight the parts still don't looks great. Could you please give some pointers as to what could be improved? Would the fact that some shoe images are cropped at the corners affect the training? I see that the semantic loss doesn't really go down and am wondering why that would be.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ 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. ๐๐๐
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 โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.