zerchen / gsdf Goto Github PK

Hi, I am writing to inquire about the availability of pre-trained weights for gSDF. I have been following your project with great interest and am keen to explore its capabilities. Thank you for your time and effort in developing this project. I look forward to your response.

Excellent work! I have a question: you mentioned in the paper that the obman dataset ran for 60 hours on 4x3090 with a batch_size of 256. However, when I followed the training steps you outlined in the readme and ran your code on 2x3090, I found it to be quite slow: 1 epoch took 1 hour with my batch_size set to 128 (due to memory constraints). Do you know why this might be?

@zerchen When I try to download the sdf.tar.gz file from here, I get:

Too many users have viewed or downloaded this file recently. Please try accessing the file again later. 
If the file you are trying to access is particularly large or is shared with many people, it may take up to 
24 hours to be able to view or download the file. If you still can't access a file after 24 hours, contact 
your domain administrator. 

Can you please check if you could re-configure the download or share the file somehow differently?

Thank you!

Hello, I use your preprocess/cocoify_obman.py and generate lmdb file, but when I train the pose_kpt task, it seems there's some error of imdb file. Speficically, It seems that some images are written to NoneType in lmdb file. Could you please help me solve this problem, since I try to find some bugs but failed?

Hello! Thanks for the great work! I'm now implementing another dataset for training, and have got some trouble making the sdf data. Could you tell me how you created the SDF data for obman and dexycb datasets? Thank you again!

Hi,

This is a very cool work.
Wondering how to run your method to test its performance of 3D hand-object reconstruction on some random images?
Do you have demo code and model weights available for quick testing?

@zerchen nice work!
You provide the sdf files for download. How were these files created?

Hello, Can you give the public access to the dexycb json file?

Thank you.