Research map / Image generation / Fast diffusion sampling
ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| VSD instead of SDS (ablation) | NeRF text-to-3D · elephant skull prompt | qualitative visual quality | Helped |
| increase NeRF rendering resolution 64 -> 512 (ablation) | SDS NeRF training · elephant skull prompt | qualitative visual quality | Helped |
| add annealed time-step schedule (ablation) | NeRF training · elephant skull prompt | qualitative visual quality | Helped |
| add scene initialization (hollow density) (ablation) | large scene generation · smart home prompt | 360 scene coverage | Helped |
| VSD instead of SDS for mesh texture fine-tuning (ablation) | ProlificDreamer mesh stage · elephant skull prompt | qualitative visual quality | Helped |
| increase number of particles 1 -> 8 (ablation) | VSD · ice cream sundae prompt | diversity | Helped |
| increase number of particles 1 -> 8 (ablation) | VSD · ice cream sundae prompt | quality | No clear effect |
| small CFG (7.5) (ablation) | SDS · text-to-3D | plausibility | Hurt |
| LoRA of pretrained model instead of small U-Net for score model | VSD · text-to-3D | fidelity | Helped |
Closest papers by meaning
- Consistent3D: Towards Consistent High-Fidelity Text-to-3D Generation with Deterministic Sampling Prior (CVPR 2024 · similarity 0.93)
- Walking the Schrödinger Bridge: A Direct Trajectory for Text-to-3D Generation (NeurIPS 2025 · similarity 0.91)
- Consistent Flow Distillation for Text-to-3D Generation (ICLR 2025 · similarity 0.91)
- Taming Mode Collapse in Score Distillation for Text-to-3D Generation (CVPR 2024 · similarity 0.91)
- DreamTime: An Improved Optimization Strategy for Diffusion-Guided 3D Generation (ICLR 2024 · similarity 0.91)
- Rethinking Score Distilling Sampling for 3D Editing and Generation (ICML 2025 · similarity 0.91)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.