andresprados / SPIGA

SPIGA: Shape Preserving Facial Landmarks with Graph Attention Networks.
https://bmvc2022.mpi-inf.mpg.de/155/
BSD 3-Clause "New" or "Revised" License
299 stars 32 forks source link

Code release #1

Open andresprados opened 2 years ago

andresprados commented 2 years ago

Code release will be available within the next two weeks, we expect to release the inference and the realtime demo before November and the training source before December.

Shubhamai commented 1 year ago

Amazing! Looking forward to it!

ucalyptus2 commented 1 year ago

@andresprados any updates?

ofirkris commented 1 year ago

@andresprados is there a change in plan of code release?

andresprados commented 1 year ago

Sorry for the delay, we hope to have it ready before the start of the conference (21/11/2022).

Shubhamai commented 1 year ago

Thanks a lot for the update :)

Oguzhanercan commented 1 year ago

How can I run the inference code?

nishitanand commented 1 year ago

Hello, thank you for the amazing work. When will the training code be released? And how can I run the inference code? I tried but wasn't able to. The code doesn't seem complete. I tried running framework.py, config.py and pretreatment.py in the inference folder, but it didn't work.

andresprados commented 1 year ago

Hi, the demo with an inference example is expected to be released next week (01/02/2023). In the meantime, the readme shows how to generate results from different datasets in the evaluation section.

lucasjinreal commented 1 year ago

@andresprados Hello, does the 3d pose estimated also sota compare with previously methods? How's the accuracy and speed tradeoff? Can it in realtime on CPU?

andresprados commented 1 year ago

Hi, the demo with an inference example is expected to be released next week (01/02/2023). In the meantime, the readme shows how to generate results from different datasets in the evaluation section.

Inference framework and image colab demo are already released! Next week I will add a real-time video demo that will include:

I share with you the preview:

Sparno1179 commented 1 year ago

Hi, @andresprados

Thank you for sharing your wonderful project. I'm very impressed with your work.

I have noticed that the training code has not yet been released, and I was wondering if you could kindly let me know when it is scheduled to be made available. I am very keen to explore and experiment with SPIGA, and having access to the training code would be helpful.

Thank you.

andresprados commented 1 year ago

Hi @Sparno1179, thank you for the support.

At the moment, the training code is integrated into a larger framework supported by our lab. I believe that in the next 2-3 months I should be able to create a simple training process for SPIGA. In the meantime, check out the paper supplementary, a detailed explanation of the training process as well as the hyperparameters used can be found there. Also take a look at the dataloaders in the repository, we have included all the data augmentors used as well as the configuration set during the training.

Best, Andrés

Kaktusava commented 1 year ago

Hi @andresprados, thank you for your amazing work!

Do you have any updates on the code release for training? Or maybe you can post the training process as it currently is?

Xiejiahao233 commented 1 year ago

Hi @andresprados, thank you for your excellent work, I am very interested in this. When will the training code be released? Or I see you said that the training code had been integrated into the framework of your lab. Can you provide a link? Thank you.