https://github.com/mosure/bevy_zeroverse
bevy zeroverse synthetic dataset generator
https://github.com/mosure/bevy_zeroverse
bevy reconstruction synthetic-dataset-generation zeroverse
Last synced: 8 months ago
JSON representation
bevy zeroverse synthetic dataset generator
- Host: GitHub
- URL: https://github.com/mosure/bevy_zeroverse
- Owner: mosure
- License: mit
- Created: 2024-06-24T23:54:22.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-10-28T16:55:11.000Z (11 months ago)
- Last Synced: 2024-10-29T21:01:32.261Z (11 months ago)
- Topics: bevy, reconstruction, synthetic-dataset-generation, zeroverse
- Language: Rust
- Homepage: https://mosure.github.io/bevy_zeroverse/project
- Size: 170 MB
- Stars: 5
- Watchers: 2
- Forks: 0
- Open Issues: 21
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# bevy_zeroverse ♾️
[](https://github.com/Mosure/bevy_zeroverse/actions?query=workflow%3Atest)
[](https://raw.githubusercontent.com/mosure/bevy_zeroverse/main/LICENSE)
[](https://github.com/mosure/bevy_zeroverse)
[](https://github.com/mosure/bevy_zeroverse/issues)
[](http://isitmaintained.com/project/mosure/bevy_zeroverse)
[](https://crates.io/crates/bevy_zeroverse)### [arXiv](https://arxiv.org/abs/) | [project page](https://mosure.github.io/bevy_zeroverse/project/index.html)
bevy zeroverse synthetic reconstruction dataset generator. view the [live demo](https://mosure.github.io/bevy_zeroverse?yaw_speed=0.7&cameras_x=2&cameras_y=2®enerate_ms=8000&plucker_visualization=true).
## capabilities
- [X] depth/normal rendering modes
- [X] plücker camera labels
- [X] generate parameteric zeroverse primitives
- [X] primitive deformation
- [x] procedural zeroverse composite environments
- [x] online torch dataloader
- [x] safetensor chunking
- [ ] primitive boolean operations
- [ ] primitive pbr wireframe
- [ ] primitive 4d augmentation## dataloader

```python
from bevy_zeroverse_dataloader import BevyZeroverseDataset
from torch.utils.data import DataLoaderdataset = BevyZeroverseDataset(
editor=False, headless=True, num_cameras=6,
width=640, height=480, num_samples=1e6,
)
dataloader = DataLoader(
dataset, batch_size=4, shuffle=True, num_workers=1,
)for batch in dataloader:
visualize(batch)
```## mat-synth
- download the mat-synth dataset [here](https://huggingface.co/datasets/gvecchio/MatSynth/blob/main/scripts/download_dataset.py)
- resize the mat-synth dataset (4k is heavy) using `python mat-synth/resize.py --source_dir --dest_dir assets/materials`
- material basecolor grid view (`cargo run -- --material-grid` or [live demo](https://mosure.github.io/bevy_zeroverse?material_grid=true))
## compatible bevy versions
| `bevy_zeroverse` | `bevy` |
| :-- | :-- |
| `0.2` | `0.14` |
| `0.1` | `0.13` |## credits
- [lgm](https://github.com/3DTopia/LGM)
- [mat-synth](https://huggingface.co/datasets/gvecchio/MatSynth)
- [zeroverse](https://github.com/desaixie/zeroverse)