Pith. sign in

REVIEW 3 major objections 3 minor 57 references

Reconstruction-Anchored Diffusion Model for Text-to-Motion Generation

T0 review · 3 major / 3 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read RAM improves diffusion-based text-to-motion generation by anchoring the denoising process to a motion-centric latent space and using reconstruction errors as guidance, achieving state-of-the-art FID and semantic alignment on HumanML3D.

desk verdict The two-stream training and ablations are solid, but the headline FID gain from REG is not cleanly attributed — the paper never tests CFG at matched guidance strength — so the empirical claim is conditional until that's fixed. read the letter →

arxiv 2601.14788 v2 pith:YUWQO52W submitted 2026-01-21 cs.CV

Yifei Liu , Changxing Ding , Ling Guo , Huaiguang Jiang , Qiong Cao This is my paper
classification cs.CV
keywords text-to-motiongenerationdiffusionmodelsmotionlatentspacereconstructionguidanceerrorpropagationself-regularizationalignmentclassifier-free
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that two key weaknesses of diffusion-based text-to-motion models—the gap between pre-trained text embeddings and actual motion dynamics, and error accumulation during iterative denoising—can both be addressed by leveraging a motion reconstruction branch. RAM co-trains this branch with the diffusion decoder, then at inference uses it to reconstruct the previous denoising step's output and amplifies the difference between the current text-driven prediction and that reconstruction. The claimed result is a large improvement in motion realism (FID 0.032 on HumanML3D) and semantic alignment (R-Precision@1 56.1%) with only 20 sampling steps, placing diffusion methods on par with or ahead of VQ-VAE-based generators. A sympathetic reader would care because it suggests diffusion models do not have to sacrifice fidelity for flexibility.

What carries the argument

The central object is the motion-centric latent space produced by a transformer motion encoder Em, together with the Reconstructive Error Guidance (REG) term. The diffusion decoder is trained to both reconstruct motion from Em's latent and generate from text latents, while two objectives shape the space: self-regularization, a softmax cross-entropy that pushes different motion latents apart, and motion-centric latent alignment, which pulls text latents toward motion latents while almost fully stop-gradients on Em. At inference, the final prediction is x̂t,s = D(xt,t,zt) + w1(D(xt,t,zt) − D(xt,t,Em(x̂t+1,s))) plus the CFG term, where the reconstruction of the previous estimate acts as a delib

What would settle it

On a held-out set of real motions, compute the cosine similarity between the REG residual (D(xt,t,zt) − D(xt,t,Em(x̂t+1,s))) and the true error direction (x0 − x̂t,s); if the average similarity is not positive across early steps, REG is not correcting toward the data. Alternatively, run the same 20-step sampler with REG replaced by a residual computed from a randomly reconstructed latent: if FID remains at 0.032, the specific motion encoder is not the cause.

Watch

Extended reading notes

Core claim

RAM claims that a motion-centric latent space, learned by reconstructing motion sequences and regularized by self-regularization and motion-centric latent alignment, gives text embeddings a target space that carries kinematic structure. At sampling time, passing the previous step's estimate through the same motion encoder and subtracting the resulting prediction from the current text-conditioned prediction yields a residual that, when amplified, steers the sample back toward clean motion. On HumanML3D, the combined system reports FID 0.032 and R-Precision@1 56.1% with 20 inference steps, surpassing previous diffusion methods and outperforming most VQ-VAE-based models.

Load-bearing premise

REG assumes that the motion encoder, trained only on clean motion sequences, produces a useful 'weakened' prediction when fed an off-manifold previous estimate, such that the amplified residual points toward clean data rather than adding noise.

Editorial extensions

If this is right

  • Diffusion-based text-to-motion models can reach the FID levels of VQ-VAE models, narrowing the historical fidelity gap.
  • Applying REG only in early denoising steps already yields most of the FID gain, confirming that error propagation is dominated by early steps.
  • REG and classifier-free guidance are complementary: CFG improves text-motion alignment while REG improves realism, and both can be combined.
  • The model is efficient: with 20 inference steps and REG enabled, average inference time per sentence remains below a standard 50-step diffusion baseline.
  • Motion-centric alignment with a nearly frozen motion encoder outperforms bidirectional alignment and contrastive strategies on realism, while matching them on semantic metrics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • REG is, in principle, a training-free plug-in for any two-stream diffusion model that already has a reconstruction encoder; it could be lifted to other generation domains, provided the encoder is calibrated on off-manifold inputs.
  • The paper's ablations leave open whether REG's benefit comes from true error correction or from a generic sharpening of the prediction; a direct test would compare it against adding a random or unconditional residual of the same magnitude.
  • Because the motion encoder is only trained on clean motion, REG's premise is fragile: if it is not robust to off-manifold estimates, the amplified residual could inject noise; a testable extension is to also train the encoder on denoised or intermediate samples.
  • The self-regularization loss likely increases semantic resolution of the latent space, so a natural extension is to evaluate the same learned latents on motion-text retrieval tasks.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 3 minor

Summary. The paper proposes Reconstruction-Anchored Diffusion Model (RAM) for text-to-motion generation. The method has two main components: (1) a motion reconstruction branch that co-trains a motion encoder and a diffusion decoder with self-regularization and motion-centric latent alignment, aiming to map text embeddings into a motion-sensitive latent space; and (2) Reconstructive Error Guidance (REG), an inference-time mechanism that reconstructs the previous denoising estimate with the motion encoder and uses the amplified residual between the current text-driven prediction and this reconstruction to correct error propagation. On HumanML3D, the authors report an FID of 0.032 and R-Precision@1 of 56.1% with 20 inference steps, claiming state-of-the-art performance among diffusion-based methods and competitiveness with VQ-VAE-based models; consistent but more modest gains are reported on KIT-ML.

Significance. If the reported results are robust, the paper makes a meaningful empirical contribution: it demonstrates that a diffusion-based text-to-motion model can close the FID gap with VQ-VAE-based models, and it proposes a novel guidance mechanism that is conceptually interesting. The paper is thorough in its ablation coverage—component ablations, latent-space training variants, hyperparameter sensitivity, inference-efficiency experiments, and a user study—and the benchmark numbers are internally consistent. However, the central causal claim for REG is not convincingly isolated: the reported improvement may be attributable to guidance-strength effects or test-set overfitting rather than the proposed error-propagation-correction mechanism. Because this is the paper's headline contribution, the manuscript currently overclaims its evidence.

major comments (3)
  1. [Section 3.3, Eq. (11), Table 5] The ablation does not isolate REG from guidance-strength effects. Eq. (11) can be rewritten as (1+w1+w2)D(xt,t,zt) - w1 D(xt,t,zm,t+1) - w2 D(xt,t,∅). With w1=5 and w2=1.5, the conditional term is scaled by 7.5, far beyond any CFG-only value reported (the largest is w2=4.5). Table 5 shows w1=5 alone gives FID 0.128 and w2=1.5 alone gives 0.132, yet their combination gives 0.032. A CFG-only control at matched effective guidance strength (e.g., w2=6 or 7.5) is missing. Without this control, the FID gain attributed to REG could be an artifact of the large combined extrapolation rather than the proposed error-propagation correction.
  2. [Section 3.3, Eq. (10), Section 3.1, Eq. (1)] REG assumes that Em(sm, xhat_{t+1,s}) is a meaningful encoding of an off-manifold estimate and that the residual D(xt,t,zt) - D(xt,t,zm,t+1) points toward clean data. However, Em is trained only on clean motions (Eq. (1)); no evidence is provided that Em generalizes to the noisy/off-manifold inputs encountered during sampling. The mechanism is supported only by the ablation in Table 5. Please add diagnostics: compare the distribution of zm,t+1 with training-time latents, measure the correlation between the REG residual and the clean-data direction, and include a control where zm,t+1 is replaced by a random or perturbed latent. Without such analysis, the claim that REG corrects error propagation is not established.
  3. [Sections 4.2, 4.5, 4.7; Tables 5, 7, 8, 9] The hyperparameters w1, w2, β, τ, w_sr, w_latent, and d_E are selected based on test-set metrics, and no validation split is described. Since the headline FID of 0.032 is the result of these sweeps, the reported numbers may reflect test-set overfitting rather than generalizable performance. Please specify the validation protocol used for model selection, or re-report the results with a clear train/validation/test split. At minimum, show that the selected hyperparameters are stable on a held-out portion of the data.
minor comments (3)
  1. [Section 3.2, Eq. (7)] The text says 'treating only identical indices as positive pairs,' but the formula is actually a uniformity loss: the numerator exp(sim(zi,zi)/τ) is constant (sim=1 after normalization), and the loss pushes all latents apart. Please revise the wording to avoid implying a supervised positive-pair contrastive loss.
  2. [Table 3 caption and table formatting] There are typos such as 'inital' (should be 'initial'), and several table captions have 'T able' instead of 'Table'. Please proofread.
  3. [Section 6 and Abstract] The code is only promised ('will be released'). Given that REG is an inference-time heuristic whose exact implementation matters, please include a pseudocode algorithm for the inference loop and, ideally, release the code and checkpoints to support reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: RAM's losses and REG update are not defined in terms of the headline metrics, and no load-bearing self-citation is used.

full rationale

The paper's derivation chain consists of reconstruction/generation regression losses (Eqs. 3–6), a self-regularization cross-entropy loss (Eq. 7), a motion-centric latent alignment loss (Eq. 8), and an inference-time Reconstructive Error Guidance update (Eqs. 10–11). None of these quantities is defined in terms of the headline FID or R-Precision values; those values are computed with the standard external HumanML3D/KIT-ML evaluation protocol, not with RAM's own latent space. REG is an additive sampling update, not a parameter fitted to data and then reported as a prediction. The reconstruction branch is trained on clean/noisy motion, and applying it to off-manifold previous estimates is an extrapolation—this is a correctness/robustness risk, not circularity. The two-stream architecture is explicitly attributed to external prior work (Language2Pose [18], TEMOS [42]); the encoders are taken from TEMOS and the decoder from MDM [21], with no authorial uniqueness theorem or self-citation chain carrying the central claim. The related-work section positions REG relative to external weak-score guidance methods [46, 49]. Hyperparameter sweeps in Tables 5, 7–9 are empirical model selection on the test set; this raises overfitting/soundness concerns but does not reduce the claimed results to their inputs by construction. The only limitation passage—future evaluation on noisy pseudo-labeled data—does not assert or imply circularity. Therefore, no quotable equation-level circular step exists, and the appropriate score is 0.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

No new physical or ontological entities are introduced; the "motion-centric latent manifold" is a learned representation, not an invented entity. The load-bearing assumptions are the standard diffusion formalism, the suitability of pretrained encoders, and the unproven REG self-correction premise.

free parameters (8)
  • beta (latent alignment gradient control) = 0.01
    Controls gradient flow to the motion encoder in Llatent (Eq. 8); chosen by the Table 7 sweep.
  • tau (self-regularization temperature) = 1.0
    Sharpness of cosine similarities in Lsr (Eq. 7); chosen by the Table 7 sweep.
  • w_sr (self-regularization loss weight) = 1.0
    Weight of the self-regularization objective; selected via the Table 8 sweep.
  • w_latent (latent alignment loss weight) = 0.5
    Weight of motion-centric latent alignment; selected via the Table 8 sweep.
  • w1 (REG amplification weight) = 5.0
    Amplification strength in Eq. 10-11; selected via the Table 5 sweep.
  • w2 (CFG weight) = 1.5
    Classifier-free guidance strength; selected via the Table 5 sweep.
  • d_E (encoder latent dimension) = 256
    Latent size for HumanML3D; selected via the Table 9 sweep.
  • inference step schedule = 20 linear-spaced steps [49,46,...,0]
    Inference efficiency choice; Table 3 compares against 50-step MDM.
assumptions (5)
  • standard math The MDM forward/reverse Gaussian diffusion formalism correctly models the motion data distribution.
    Section 3.1; standard diffusion assumption, not proven in the paper.
  • domain assumption Using the TEMOS text/motion encoders and DistilBERT embeddings provides a sufficiently informative starting representation.
    Section 4.2; relies on upstream pretrained models.
  • domain assumption A latent space learned by motion reconstruction is the right intermediate supervision for text-to-motion generation, and Lsr improves its semantic resolution.
    Section 3.2; central design hypothesis supported only by ablations.
  • ad hoc to paper At inference, D(xt,t,Em(x_hat_{t+1,s})) reproduces the previous step's error patterns and the amplified residual points toward clean data (REG).
    Section 3.3, Eq. 10-11; no derivation, only empirical justification.
  • ad hoc to paper Setting beta=0.01 with stop-gradient keeps the latent space motion-centric while still allowing alignment to converge.
    Section 3.2 and Table 7; empirical design choice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Reconstruction-Anchored Diffusion Model for Text-to-Motion Generation." pith.science (2026). https://pith.science/paper/YUWQO52W

@misc{pith2026260114788,
  author       = {Pith},
  title        = {Pith review of: Reconstruction-Anchored Diffusion Model for Text-to-Motion Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YUWQO52W}},
  note         = {Machine review of arXiv:2601.14788}
}
read the original abstract

Diffusion models have seen widespread adoption for text-driven human motion generation and related tasks due to their impressive generative capabilities and flexibility. However, current motion diffusion models face two major limitations: a representational gap caused by pre-trained text encoders that lack motion-specific information, and error propagation during the iterative denoising process. This paper introduces Reconstruction-Anchored Diffusion Model (RAM) to address these challenges. First, RAM leverages a motion latent space as intermediate supervision for text-to-motion generation. To this end, RAM co-trains a motion reconstruction branch with two key objective functions: self-regularization to enhance the discrimination of the motion space and motion-centric latent alignment to enable accurate mapping from text to the motion latent space. Second, we propose Reconstructive Error Guidance (REG), a testing-stage guidance mechanism that exploits the motion diffusion model's inherent self-correction ability to mitigate error propagation. At each denoising step, REG uses the motion reconstruction branch to reconstruct the previous estimate, reproducing the prior error patterns. By amplifying the residual between the current prediction and the reconstructed estimate, REG highlights the improvements in the current prediction. Extensive experiments demonstrate that RAM achieves significant improvements and state-of-the-art performance. Our code will be released.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

57 extracted references · 9 linked inside Pith

  1. [1]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Du, Y., Kips, R., Pumarola, A., Starke, S., Thabet, A., Sanakoyeu, A.: Avatars grow legs: Generating smooth human motion from sparse tracking inputs with diffusion model. In: Computer Vision and Pattern Recognition (CVPR), pp. 481– 490 (2023)

    2023

  2. [2]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Liang, H., Bao, J., Zhang, R., Ren, S., Xu, Y., Yang, S., Chen, X., Yu, J., Xu, L.: Omg: Towards open-vocabulary motion generation via mixture of controllers. In: Computer Vision and Pattern Recognition (CVPR), pp. 482–493 (2024)

    2024

  3. [3]

    18 In: International Conference on Robotics and Automation (ICRA), pp

    Xia, F., Li, C., Mart ´ ın-Mart ´ ın, R., Litany, O., Toshev, A., Savarese, S.: Relmogen: Integrating motion generation in reinforcement learning for mobile manipulation. 18 In: International Conference on Robotics and Automation (ICRA), pp. 4583–4590 (2021). IEEE

    2021

  4. [4]

    In: International Conference on Computer Graphics and Interactive Techniques (SIGGRAPH), pp

    Cohan, S., Tevet, G., Reda, D., Peng, X.B., Panne, M.: Flexible motion in- betweening with diffusion models. In: International Conference on Computer Graphics and Interactive Techniques (SIGGRAPH), pp. 1–9 (2024)

    2024

  5. [5]

    In: European Conference on Computer Vision (ECCV), pp

    Li, J., Clegg, A., Mottaghi, R., Wu, J., Puig, X., Liu, C.K.: Controllable human-object interaction synthesis. In: European Conference on Computer Vision (ECCV), pp. 54–72 (2024). Springer

    2024

  6. [6]

    International Journal of Computer Vision (IJCV), 1–21 (2024)

    Liang, H., Zhang, W., Li, W., Yu, J., Xu, L.: Intergen: Diffusion-based multi- human motion generation under complex interactions. International Journal of Computer Vision (IJCV), 1–21 (2024)

    2024

  7. [7]

    International Journal of Computer Vision (IJCV)133(5), 2910–2926 (2025)

    Sun, M., Xu, C., Jiang, X., Liu, Y., Sun, B., Huang, R.: Beyond talking–generating holistic 3d human dyadic motion for communication. International Journal of Computer Vision (IJCV)133(5), 2910–2926 (2025)

    2025

  8. [8]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Wang, Y., Wang, S., Zhang, J., Fan, K., Wu, J., Xue, Z., Liu, Y.: Timotion: Tem- poral and interactive framework for efficient human-human motion generation. In: Computer Vision and Pattern Recognition (CVPR), pp. 7169–7178 (2025)

    2025

Show all 57 references
  1. [9]

    International Journal of Computer Vision (IJCV)134(1), 61 (2026) https://doi.org/10.1007/ s11263-025-02611-3

    Dai, Y., Zhu, W., Li, R.,et al.: Tcdiff++: An end-to-end trajectory-controllable diffusion model for harmonious music-driven group choreography. International Journal of Computer Vision (IJCV)134(1), 61 (2026) https://doi.org/10.1007/ s11263-025-02611-3

    2026

  2. [10]

    arXiv preprint arXiv:2108.08877 (2021)

    Ni, J., Abrego, G.H., Constant, N., Ma, J., Hall, K.B., Cer, D., Yang, Y.: Sentence-t5: Scalable sentence encoders from pre-trained text-to-text models. arXiv preprint arXiv:2108.08877 (2021)

    2021 arXiv

  3. [11]

    In: International Conference on Machine Learning (ICML), pp

    Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J.,et al.: Learning transferable visual models from natural language supervision. In: International Conference on Machine Learning (ICML), pp. 8748–8763 (2021). PmLR

    2021

  4. [12]

    arXiv preprint arXiv:1910.01108 (2019)

    Sanh, V., Debut, L., Chaumond, J., Wolf, T.: Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter. arXiv preprint arXiv:1910.01108 (2019)

    1910 arXiv

  5. [13]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Chen, X., Jiang, B., Liu, W., Huang, Z., Fu, B., Chen, T., Yu, G.: Executing your commands via motion diffusion in latent space. In: Computer Vision and Pattern Recognition (CVPR), pp. 18000–18010 (2023)

    2023

  6. [14]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Sheng, J., Lin, M., Zhao, A., Pruvost, K., Wen, Y.-H., Li, Y., Huang, G., Liu, Y.-J.: Exploring text-to-motion generation with human preference. In: Computer Vision and Pattern Recognition (CVPR), pp. 1888–1899 (2024) 19

    2024

  7. [15]

    Advances in Neural Information Processing Systems (NeurIPS)36, 15497–15518 (2023)

    Jin, P., Wu, Y., Fan, Y., Sun, Z., Yang, W., Yuan, L.: Act as you wish: Fine- grained control of motion diffusion model with hierarchical semantic graphs. Advances in Neural Information Processing Systems (NeurIPS)36, 15497–15518 (2023)

    2023

  8. [16]

    In: International Conference on Computer Vision (ICCV), pp

    Zhang, M., Guo, X., Pan, L., Cai, Z., Hong, F., Li, H., Yang, L., Liu, Z.: Remod- iffuse: Retrieval-augmented motion diffusion model. In: International Conference on Computer Vision (ICCV), pp. 364–373 (2023)

    2023

  9. [17]

    In: International Conference on Learning Representations (ICLR) (2024)

    Li, Y., Schaar, M.: On error propagation of diffusion models. In: International Conference on Learning Representations (ICLR) (2024)

    2024

  10. [18]

    In: International Conference on 3D Vision (3DV), pp

    Ahuja, C., Morency, L.-P.: Language2pose: Natural language grounded pose fore- casting. In: International Conference on 3D Vision (3DV), pp. 719–728 (2019). IEEE

    2019

  11. [19]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Guo, C., Zou, S., Zuo, X., Wang, S., Ji, W., Li, X., Cheng, L.: Generating diverse and natural 3d human motions from text. In: Computer Vision and Pattern Recognition (CVPR), pp. 5152–5161 (2022)

    2022

  12. [20]

    Big data 4(4), 236–252 (2016)

    Plappert, M., Mandery, C., Asfour, T.: The kit motion-language dataset. Big data 4(4), 236–252 (2016)

    2016

  13. [21]

    arXiv preprint arXiv:2209.14916 (2022)

    Tevet, G., Raab, S., Gordon, B., Shafir, Y., Cohen-Or, D., Bermano, A.H.: Human motion diffusion model. arXiv preprint arXiv:2209.14916 (2022)

    2022 arXiv

  14. [22]

    arXiv preprint arXiv:2208.15001 (2022)

    Zhang, M., Cai, Z., Pan, L., Hong, F., Guo, X., Yang, L., Liu, Z.: Motiondif- fuse: Text-driven human motion generation with diffusion model. arXiv preprint arXiv:2208.15001 (2022)

    2022 arXiv

  15. [23]

    Advances in Neural Information Processing Systems (NeurIPS)36, 13981–13992 (2023)

    Zhang, M., Li, H., Cai, Z., Ren, J., Yang, L., Liu, Z.: Finemogen: Fine-grained spatio-temporal motion generation and editing. Advances in Neural Information Processing Systems (NeurIPS)36, 13981–13992 (2023)

    2023

  16. [24]

    arXiv preprint arXiv:2503.13859 (2025)

    Bae, J., Hwang, I., Lee, Y.Y., Guo, Z., Liu, J., Ben-Shabat, Y., Kim, Y.M., Kapa- dia, M.: Less is more: Improving motion diffusion models with sparse keyframes. arXiv preprint arXiv:2503.13859 (2025)

    2025 arXiv

  17. [25]

    In: European Conference on Computer Vision (ECCV), pp

    Dai, W., Chen, L.-H., Wang, J., Liu, J., Dai, B., Tang, Y.: Motionlcm: Real- time controllable motion generation via latent consistency model. In: European Conference on Computer Vision (ECCV), pp. 390–408 (2024). Springer

    2024

  18. [26]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Hong, S., Kim, C., Yoon, S., Nam, J., Cha, S., Noh, J.: Salad: Skeleton-aware latent diffusion for text-driven motion generation and editing. In: Computer Vision and Pattern Recognition (CVPR), pp. 7158–7168 (2025)

    2025

  19. [27]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Zhang, J., Zhang, Y., Cun, X., Zhang, Y., Zhao, H., Lu, H., Shen, X., Shan, Y.: 20 Generating human motion from textual descriptions with discrete representations. In: Computer Vision and Pattern Recognition (CVPR), pp. 14730–14740 (2023)

    2023

  20. [28]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Pinyoanuntapong, E., Wang, P., Lee, M., Chen, C.: Mmm: Generative masked motion model. In: Computer Vision and Pattern Recognition (CVPR), pp. 1546– 1555 (2024)

    2024

  21. [29]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Guo, C., Mu, Y., Javed, M.G., Wang, S., Cheng, L.: Momask: Generative masked modeling of 3d human motions. In: Computer Vision and Pattern Recognition (CVPR), pp. 1900–1910 (2024)

    1900

  22. [30]

    In: European Conference on Computer Vision (ECCV), pp

    Pinyoanuntapong, E., Saleem, M.U., Wang, P., Lee, M., Das, S., Chen, C.: Bamm: Bidirectional autoregressive motion model. In: European Conference on Computer Vision (ECCV), pp. 172–190 (2024). Springer

    2024

  23. [31]

    Advances in Neural Information Processing Systems (NeurIPS)37, 130739–130763 (2024)

    Yuan, W., He, Y., Shen, W., Dong, Y., Gu, X., Dong, Z., Bo, L., Huang, Q.: Mogents: Motion generation based on spatial-temporal joint modeling. Advances in Neural Information Processing Systems (NeurIPS)37, 130739–130763 (2024)

    2024

  24. [32]

    In: International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp

    Hosseyni, S.R., Rahmani, A.A., Seyedmohammadi, S.J., Seyedin, S., Mohammadi, A.: Bad: Bidirectional auto-regressive diffusion for text-to-motion generation. In: International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1–5 (2025). IEEE

    2025

  25. [33]

    International Conference on Computer Vision (ICCV) (2025)

    Zhang, P., Liu, P., Kim, H., Garrido, P., Chaudhuri, B.: Kinmo: Kinematic- aware human motion understanding and generation. International Conference on Computer Vision (ICCV) (2025)

    2025

  26. [34]

    International Conference on Learning Representations (ICLR) (2025)

    Li, Z., Yuan, W., He, Y., Qiu, L., Zhu, S., Gu, X., Shen, W., Dong, Y., Dong, Z., Yang, L.T.: Lamp: Language-motion pretraining for motion generation, retrieval, and captioning. International Conference on Learning Representations (ICLR) (2025)

    2025

  27. [35]

    In: AAAI Conference on Artificial Intelligence (2026)

    Shen, K., Wu, B., Chen, J., Wang, X., Shen, L.: Finextrol: Controllable motion generation via fine-grained text. In: AAAI Conference on Artificial Intelligence (2026)

    2026

  28. [36]

    In: International Conference on Computer Graphics and Interactive Techniques in Asia (SIGGRAPH Asia), pp

    Nguyen, Q., Le, T., Huang, B., Vu, M.N., Le, N., Vo, T., Nguyen, A.: Learning human motion with temporally conditional mamba. In: International Conference on Computer Graphics and Interactive Techniques in Asia (SIGGRAPH Asia), pp. 1–10 (2025)

    2025

  29. [37]

    Conference on Neural Information Processing Systems (NeurIPS) (2025)

    Au, H.Y., Chen, J., Jiang, J., Xiang, J.: Deep compositional phase diffusion for long motion sequence generation. Conference on Neural Information Processing Systems (NeurIPS) (2025)

    2025

  30. [38]

    International Journal of Computer Vision (IJCV)134(1), 29 (2026)

    Cao, Y., Guo, X., Zhang, M., Xie, H., Gu, C., Liu, Z.: Crowdmogen: Event-driven 21 collective human motion generation. International Journal of Computer Vision (IJCV)134(1), 29 (2026)

    2026

  31. [39]

    In: European Conference on Computer Vision (ECCV), pp

    Tevet, G., Gordon, B., Hertz, A., Bermano, A.H., Cohen-Or, D.: Motionclip: Exposing human motion generation to clip space. In: European Conference on Computer Vision (ECCV), pp. 358–374 (2022). Springer

    2022

  32. [40]

    In: Computer Vision and Pattern Recognition (CVPR), pp

    Maldonado, G., Pazho, A.D., Noghre, G.A., Katariya, V., Tabkhi, H.: Moclip motion-aware fine-tuning and distillation of clip for human motion generation. In: Computer Vision and Pattern Recognition (CVPR), pp. 2931–2941 (2025)

    2025

  33. [41]

    In: International Conference on Computer Vision (ICCV), pp

    Ghosh, A., Cheema, N., Oguz, C., Theobalt, C., Slusallek, P.: Synthesis of com- positional animations from textual descriptions. In: International Conference on Computer Vision (ICCV), pp. 1396–1406 (2021)

    2021

  34. [42]

    In: European Conference on Computer Vision (ECCV), pp

    Petrovich, M., Black, M.J., Varol, G.: Temos: Generating diverse human motions from textual descriptions. In: European Conference on Computer Vision (ECCV), pp. 480–497 (2022). Springer

    2022

  35. [43]

    In: International Conference on Computer Vision (ICCV), pp

    Petrovich, M., Black, M.J., Varol, G.: Tmr: Text-to-motion retrieval using con- trastive 3d human motion synthesis. In: International Conference on Computer Vision (ICCV), pp. 9488–9497 (2023)

    2023

  36. [44]

    Con- ference on Neural Information Processing Systems (NeurIPS)34, 8780–8794 (2021)

    Dhariwal, P., Nichol, A.: Diffusion models beat gans on image synthesis. Con- ference on Neural Information Processing Systems (NeurIPS)34, 8780–8794 (2021)

    2021

  37. [45]

    arXiv preprint arXiv:2207.12598 (2022)

    Ho, J., Salimans, T.: Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598 (2022)

    2022 arXiv

  38. [46]

    Advances in Neural Information Processing Systems (NeurIPS)37, 52996–53021 (2024)

    Karras, T., Aittala, M., Kynk¨ a¨ anniemi, T., Lehtinen, J., Aila, T., Laine, S.: Guid- ing a diffusion model with a bad version of itself. Advances in Neural Information Processing Systems (NeurIPS)37, 52996–53021 (2024)

    2024

  39. [47]

    arXiv preprint arXiv:2112.10741 (2021)

    Nichol, A., Dhariwal, P., Ramesh, A., Shyam, P., Mishkin, P., McGrew, B., Sutskever, I., Chen, M.: Glide: Towards photorealistic image generation and edit- ing with text-guided diffusion models. arXiv preprint arXiv:2112.10741 (2021)

    2021 arXiv

  40. [48]

    https://github.com/Stability-AI/sd3.5 (2024)

    Stability AI: Stable Diffusion 3.5. https://github.com/Stability-AI/sd3.5 (2024)

    2024

  41. [49]

    In: European Conference on Computer Vision (ECCV), pp

    Ahn, D., Cho, H., Min, J., Jang, W., Kim, J., Kim, S., Park, H.H., Jin, K.H., Kim, S.: Self-rectifying diffusion sampling with perturbed-attention guidance. In: European Conference on Computer Vision (ECCV), pp. 1–17 (2024). Springer

    2024

  42. [50]

    In: International Conference on Computer Vision, pp

    Mahmood, N., Ghorbani, N., Troje, N.F., Pons-Moll, G., Black, M.J.: AMASS: Archive of Motion Capture as Surface Shapes. In: International Conference on Computer Vision, pp. 5442–5451 (2019) 22

    2019

  43. [51]

    International Conference on Multimedia (MM) (2024)

    Huang, Y., Yang, H., Luo, C., Wang, Y., Xu, S., Zhang, Z., Zhang, M., Peng, J.: Stablemofusion: Towards robust and efficient diffusion-based motion generation framework. International Conference on Multimedia (MM) (2024)

    2024

  44. [52]

    In: International Conference on Learning Representations (ICLR) (2025)

    Tevet, G., Raab, S., Cohan, S., Reda, D., Luo, Z., Peng, X.B., Bermano, A.H., Panne, M.: Closd: Closing the loop between simulation and diffusion for multi- task character control. In: International Conference on Learning Representations (ICLR) (2025)

    2025

  45. [53]

    arXiv preprint arXiv:2303.01418 (2023)

    Shafir, Y., Tevet, G., Kapon, R., Bermano, A.H.: Human motion diffusion as a generative prior. arXiv preprint arXiv:2303.01418 (2023)

    2023 arXiv

  46. [54]

    arXiv preprint arXiv:2501.05098 (2025)

    Zhang, Y., Lin, J., Zeng, A., Wu, G., Lu, S., Fu, Y., Cai, Y., Zhang, R., Wang, H., Zhang, L.: Motion-x++: A large-scale multimodal 3d whole-body human motion dataset. arXiv preprint arXiv:2501.05098 (2025)

    2025 arXiv

  47. [55]

    International Journal of Computer Vision (IJCV)129(10), 2846–2864 (2021)

    Chen, X., Pang, A., Yang, W., Ma, Y., Xu, L., Yu, J.: Sportscap: Monocular 3d human motion capture and fine-grained understanding in challenging sports videos. International Journal of Computer Vision (IJCV)129(10), 2846–2864 (2021)

    2021

  48. [56]

    International Journal of Computer Vision (IJCV)130(5), 1165–1180 (2022)

    Dong, J., Shuai, Q., Sun, J., Zhang, Y., Bao, H., Zhou, X.: imocap: motion capture from internet videos. International Journal of Computer Vision (IJCV)130(5), 1165–1180 (2022)

    2022

  49. [57]

    International Journal of Computer Vision (IJCV)130(2), 363–383 (2022) 23

    Li, Z., Sedlar, J., Carpentier, J., Laptev, I., Mansard, N., Sivic, J.: Estimat- ing 3d motion and forces of human–object interactions from internet videos. International Journal of Computer Vision (IJCV)130(2), 363–383 (2022) 23

    2022

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.