zj-dong / ag3d Goto Github PK
View Code? Open in Web Editor NEWOfficial code release for ICCV2023 paper AG3D: Learning to Generate 3D Avatars from 2D Image Collections
Home Page: https://zj-dong.github.io/AG3D/
Official code release for ICCV2023 paper AG3D: Learning to Generate 3D Avatars from 2D Image Collections
Home Page: https://zj-dong.github.io/AG3D/
Thanks for sharing this great work, I'm trying to compare with your method and have to re-train your model. I saw the training folder but there are no instructions on how to run it, maybe could you provide some instructions? Thanks!
it is not clonning on command prompt and stucked by giving EOF error at 29 percent, also not working on colab when we activate and also not having pytorch exact version
Hi, where could we download the evaluation data eva3d_icon.zip
to reproduce the numbers? Thanks
Hi, thank you for sharing your work. I just want to know about does you model can do body texture?
I also want to know how did you animate these mesh model? Did you use auto rigged function ? Can we output 3D mesh obj ?
Can I get a mesh with color and texture?
Environment Information:
• OS: Ubuntu (specify version, e.g., 22.04)
• Python Version: 3.9
• PyTorch Version: 2.0.1
• CUDA Version: 11.7
• GCC Version: 11.4.0
• G++ Version: 11.4.0
• PyTorch3D Version: 0.7.6 (if applicable)
Issue Description: Encountering a TypeError when trying to run the test script test.py. The command used was:
python test.py --network=./model/deep_fashion.pkl --pose_dist=./data/dp_pose_dist.npy --output_path './result/deepfashion' --res=512 --truncation=0.7 --number=100 --type=gen_samples
The error message points to the cuda_gridsample.py file, with the specific error:
TypeError: 'tuple' object is not callable
This error occurs during the execution of a PyTorch custom CUDA operation, specifically at line 47 of cuda_gridsample.py.
Troubleshooting Steps Attempted:
The generator is pre-trained and I used the dual discriminator just like:
ag3d_discriminator = legacy.load_network_pkl(f)['D'].to(args.device)
output = G.only_synthesis(ws=ws, c=c, truncation_psi=truncation)
full_logit = ag3d_discriminator.D_image(output,c)
but i got a low confidence around -10 , is that normal?
I'm interested in your work and want to train a new model with your code. However, I don't know the contents of your datasets and the processing way of the datasets.
Do you have a schedule to tell me when the data will be released?
I am looking forward to your reply. Thank you!
What is the local identity-preserving loss of AG3D text editing ? Can it maintain the ID of the human body?
Hi, will you be releasing discriminator weights? Thanks!
May I ask which work the author obtained the 'cam2world_matrix' in the label from? Thank you very much!
can I use it without NIVIDIA GPU if yes then what is the option
Hi,
There is something it seems regarding discriminators.
I tried --
Training from Scratch
: fake_depths and fake_images look alright (similar to sampled smpl pose). As training proceeds, fake_depths become total black and white. images go full background-- (white)
Training from provided weights
: fake_depths and fake_images look alright (similar to sampled smpl pose, and realistic). As training proceeds, both deteriorate towards ---> fake_depths become total black and white. images go full background-- (white)
I speculate the reason behind it, may be a discriminator issue driving the generator crazy?
Originally posted by @shash29-dev in #14 (comment)
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.