Research map / Video / Video generation
TEMOS: Generating Diverse Human Motions from Textual Descriptions
Mathis Petrovich, Michael J. Black, Gül Varol
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 |
|---|---|---|---|
| use non-autoregressive Transformers instead of GRUs (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 1.443 -> 0.963 | Improved |
| add motion encoder (second KL loss) (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 1.178 -> 0.963 | Improved |
| add cross-modal embedding loss (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 1.091 -> 0.963 | Improved |
| add Gaussian priors (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 1.049 -> 0.963 | Improved |
| add cross-modal KL losses (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 0.993 -> 0.963 | Improved |
| make model variational (VAE sampling) instead of deterministic (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 1.175 -> 0.963 | Improved |
| sample best of 10 latent vectors instead of one random (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 0.963 -> 0.784 | Improved |
| average over 10 random samples instead of one (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 0.963 -> 1.001 | No clear effect |
| freeze DistilBERT instead of finetuning (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 1.402 -> 0.963 | Improved |
| use larger language models (BERT, RoBERTa) instead of DistilBERT (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 0.963 -> 0.986 | No clear effect |
| batch size 32 or 8 instead of 16 or 24 (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 1.260 -> 0.963 | Mixed |
| loss weights 1e-5 instead of 1e-3 (ablation) | TEMOS · KIT Motion-Language test | APE root joint: 1.219 -> 0.963 | Improved |
Later papers that compared with this method
6 later accepted papers report TEMOS as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- FineMotion: A Dataset and Benchmark with both Spatial and Temporal Annotation for Fine-grained Motion Generation and Editing (ICCV 2025 · better on 2 of 2 shared results)
- InterGen: Diffusion-Based Multi-human Motion Generation Under Complex Interactions (IJCV 2024 · better on 4 of 4 shared results)
- FLAME: Free-Form Language-Based Motion Synthesis & Editing (AAAI 2023 · better on 8 of 9 shared results)
- TMR: Text-to-Motion Retrieval Using Contrastive 3D Human Motion Synthesis (ICCV 2023 · better on 7 of 7 shared results)
- Breaking The Limits of Text-conditioned 3D Motion Synthesis with Elaborative Descriptions (ICCV 2023 · better on 5 of 5 shared results)
- Fg-T2M: Fine-Grained Text-Driven Human Motion Generation via Diffusion Model (ICCV 2023 · better on 1 of 1 shared result)
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
- HumanTOMATO: Text-aligned Whole-body Motion Generation (ICML 2024 · similarity 0.89)
- AMD: Autoregressive Motion Diffusion (AAAI 2024 · similarity 0.89)
- OmniMotionGPT: Animal Motion Generation with Limited Data (CVPR 2024 · similarity 0.89)
- SnapMoGen: Human Motion Generation from Expressive Texts (NeurIPS 2025 · similarity 0.88)
- Shape My Moves: Text-Driven Shape-Aware Synthesis of Human Motions (CVPR 2025 · similarity 0.88)
- The Language of Motion: Unifying Verbal and Non-verbal Language of 3D Human Motion (CVPR 2025 · similarity 0.88)
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.
