Image-to-Image
Safetensors
English
computer-vision
surface-normal-estimation
monocular-geometry-estimation
transparent-objects
rectified-flow
flux
lora
Instructions to use Longxiang-ai/TransNormal-2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Inference
- Notebooks
- Google Colab
- Kaggle
| { | |
| "model_name": "TransNormal-2", | |
| "task": "normal", | |
| "base_model": "black-forest-labs/FLUX.2-klein-base-9B", | |
| "lora": { | |
| "adapter_name": "core_predictor", | |
| "rank": 256, | |
| "alpha": 256 | |
| }, | |
| "grm": { | |
| "hidden_channels": 64, | |
| "edge_channels": 32, | |
| "num_blocks": 4, | |
| "use_rgb": true, | |
| "anchor_radius": 4, | |
| "anchor_eps": 0.01, | |
| "residual_scale": 0.15, | |
| "residual_clip_deg": 5.0, | |
| "initial_alpha": 0.05 | |
| }, | |
| "inference": { | |
| "single_step": true, | |
| "prompt": "", | |
| "recommended_dtype": "bfloat16" | |
| } | |
| } | |