Diffusion Generation
Cross-source consensus on Diffusion Generation from 1 sources and 5 claims.
1 sources · 5 claims
How it works
Preparation
Risks & contraindications
Comparisons
Highlighted claims
- Stable Diffusion 2.1 was fine-tuned with a rank-32 LoRA adapter for 36,000 steps. — A General Bézier Tree Encoding Counterfactual Framework for Retinal-Vessel-Mediated Disease Analysis
- VesselControlNet conditioned generation on the Bézier hint by duplicating the Stable Diffusion UNet encoder and adding zero-convolution residuals. — A General Bézier Tree Encoding Counterfactual Framework for Retinal-Vessel-Mediated Disease Analysis
- LoRA and ControlNet needed matched noise-offset co-training to avoid color drift artifacts. — A General Bézier Tree Encoding Counterfactual Framework for Retinal-Vessel-Mediated Disease Analysis
- Empty-prompt generation isolated the vessel hint more strongly than disease text prompts. — A General Bézier Tree Encoding Counterfactual Framework for Retinal-Vessel-Mediated Disease Analysis
- Higher classifier-free guidance partly recovered a hint-aligned effect, suggesting competition between text and parametric hints. — A General Bézier Tree Encoding Counterfactual Framework for Retinal-Vessel-Mediated Disease Analysis