File size: 918 Bytes
78ba666 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 |
---
license: mit
library_name: diffusers
pipeline_tag: image-to-image
---
# DiLightNet: Fine-grained Lighting Control for Diffusion-based Image Generation
SIGGRAPH 2024
- Project Page: https://dilightnet.github.io/
- Paper: https://arxiv.org/abs/2402.11929
- Full Usage: please check https://github.com/iamNCJ/DiLightNet
Example Usage:
```python
from diffusers.utils import get_class_from_dynamic_module
NeuralTextureControlNetModel = get_class_from_dynamic_module(
"dilightnet/model_helpers",
"neuraltexture_controlnet.py",
"NeuralTextureControlNetModel"
)
neuraltexture_controlnet = NeuralTextureControlNetModel.from_pretrained("DiLightNet/DiLightNet")
pipe = StableDiffusionControlNetPipeline.from_pretrained(
"stabilityai/stable-diffusion-2-1", controlnet=neuraltexture_controlnet,
)
cond_image = torch.randn((1, 16, 512, 512))
image = pipe("some text prompt", image=cond_image).images[0]
``` |