facebookresearch / keypointnerf Goto Github PK
View Code? Open in Web Editor NEWKeypointNeRF Generalizing Image-based Volumetric Avatars using Relative Spatial Encoding of Keypoints
License: Other
KeypointNeRF Generalizing Image-based Volumetric Avatars using Relative Spatial Encoding of Keypoints
License: Other
Hi, author. Thanks for your great work!
I'm trying to use your pre-trained model to render my own data, but the function "decode_batch" and the parameters "batch" seems too complicate for me. I'm wondering if there is some easy way to do it?
Look forwarding to your reply!
we can not load environment.yml,
ResolvePackageNotFound:
Hi,
Thank you for releasing the code.
I wonder how much GPU memory KeypointNeRF requires if I use more images like 16 images. Since keypointnerf needs to keep feature maps of all imput images, I want to know if 24GB GPU memory is enough for the training.
Thanks
Hi, authors
which tool are you used to generate the zju_joints3d.zip?
I am trying to train KeypointNeRF on my own dataset.
Hi, author. Thanx for your great work!
It seems that the example given is on task of reconstruction of human bodies. I wonder how to reconstruct human heads from two or three images.
Maybe I read it wrong. Look forwarding to your reply!
Is there a dataset for 3D facial reconstruction? How to generate 3d face obj files? thank you!!!
hi @markomih ,
Thank you for releasing the code.
Can you share the hardware configuration of your training?
i used a NVIDIA3090 card to run traning code, the errors are as follows:
$ nvidia-smi
Thu Jun 29 16:09:01 2023
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.105.17 Driver Version: 525.105.17 CUDA Version: 12.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 NVIDIA GeForce ... Off | 00000000:B1:00.0 Off | Off |
| 0% 41C P0 103W / 450W | 0MiB / 24564MiB | 2% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
$ python train.py --config ./configs/zju.json --data_root ./data/zju_mocap
RuntimeError: CUDA out of memory. Tried to allocate 222.00 MiB (GPU 0; 23.68 GiB total capacity; 20.55 GiB already allocated; 49.00 MiB free; 21.58 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
Hi, authors,
Thanks for your great works! I'm interesting with the performance of KeypointNeRF on the ZJU dataset. I follow the DATA_PREP to prepare dataset and start to train KeypointNeRF.
However, I had the following error:
This is caused when I built the val or test dataset. I had checked the code and found the reason:
I notice that CoreView313 and CoreView315 use different sampled cameras with other persons, see here. The number of cameras in the CoreView 313 and CoreView 315 are 21, so the index of 21 at here is out of bounds.
Could plz to check this for me?
If I have SMPL vertices, how to precalculate 3D joints?
Hi, thanks for the great work!!!
What if the 3D points out of index (let's say it may occur negative indexes of image pixel) after projected to other camera views rather than target views?
Hello, thanks for this great work!
I want to try this method on NeRF synthetic dataset (https://drive.google.com/drive/folders/1JDdLGDruGNXWnM1eqY1FNL9PlStjaKWi).
Is this method applicable to non-human data?
Hey,awesome work!
Can I using the KeypointNERF to reconstruct
a full body which can be exported as a obj file with texture.
I really appreciate the project. Unfortunately, I'm restricted by CUDA memory and limited skills. Could you provide files about the training and inferencing in multiple GPU modes?
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.