Research map / Video / Video generation
Free-Bloom: Zero-Shot Text-to-Video Generator with LLM Director and LDM Animator
Hanzhuo Huang, Yufan Feng, Cheng Shi, Lan Xu, Jingyi Yu, Sibei Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| replace joint noise initialization with independent noise (ablation) | Free-Bloom · qualitative frame consistency | content consistency | Improved |
| add step-aware self-attention shift to contextual frames (ablation) | Free-Bloom · qualitative identical/temporal coherence | coherence | Improved |
| add attention to the current frame itself (vs TAV spatio-temporal attention only) (ablation) | Free-Bloom · qualitative semantic alignment | semantic coherence | Improved |
| shift attention across denoising steps rather than concatenating current and contextual attention without shifting (ablation) | Free-Bloom · qualitative semantics | semantic coherence | Improved |
| use sin/cos weighted noise mixing instead of linear (1-lambda, lambda) weighting (ablation) | Free-Bloom joint noise · qualitative image quality | frame validity | Improved |
| add serial LLM prompts to Text2Video-Zero (ablation) | Text2Video-Zero · qualitative | semantic progression | No clear effect |
| balance dual-path interpolation coefficient m(t) (vs mostly contextual path or mostly denoising path) (ablation) | Free-Bloom interpolation · qualitative interpolation | frame quality | Improved |
| compare Free-Bloom to trained VideoFusion on temporal coherence | Free-Bloom · User study | Temporal coherence (1-5): 3.267 vs 3.889 | Hurt |
| compare Free-Bloom to T2V-Zero on fidelity | Free-Bloom · User study | Fidelity (1-5): 3.486 -> 4.133 | Improved |
| score CLIP against serial LLM prompts instead of input prompt | Free-Bloom · CLIP metrics | CLIP score: 0.477 -> 0.482 | Improved |
Later papers that compared with this method
1 later accepted paper reports Free-Bloom as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- From Prompt to Progression: Taming Video Diffusion Models for Seamless Attribute Transition (ICCV 2025 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- VideoFusion — VideoFusion: Decomposed Diffusion Models for High-Quality Video Generation (CVPR 2023) · better on 3 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- EIDT-V: Exploiting Intersections in Diffusion Trajectories for Model-Agnostic, Zero-Shot, Training-Free Text-to-Video Generation (CVPR 2025 · similarity 0.89)
- A Recipe for Scaling up Text-to-Video Generation with Text-free Videos (CVPR 2024 · similarity 0.88)
- ByTheWay: Boost Your Text-to-Video Generation Model to Higher Quality in a Training-free Way (CVPR 2025 · similarity 0.88)
- FreeNoise: Tuning-Free Longer Video Diffusion via Noise Rescheduling (ICLR 2024 · similarity 0.88)
- VideoPoet: A Large Language Model for Zero-Shot Video Generation (ICML 2024 · similarity 0.88)
- StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text (CVPR 2025 · similarity 0.87)
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.
