Text-to-motion generation has advanced with larger corpora and stronger generators, yet many models still rely on holistic frame- or clip-level latents that entangle trajectory, orientation, and articulation. This entanglement obscures body topology and forces the generator to recover kinematic structure implicitly. We present \name, a SMPL motion generation framework that factorizes motion into continuous kinematic-unit latents. A causal Motion VAE maps motion to a time-by-kinematic-unit latent manifold, and a Kinematic-Unit Flow Transformer performs text-conditioned flow matching in this structured space. Because each latent coordinate remains tied to a physical body unit, \name can use kinematic-tree rotary position encoding and kinematic-adaptive flow scheduling. We further train the generator with per-token timesteps over clean-context/noisy-target masks, enabling frame-conditioned continuation and autoregressive segment chaining within one model. Experiments first validate the representation: the kinematic-unit VAE achieves lower geometry, rotation, and feature errors than existing motion tokenizers, showing that the latent space preserves articulated structure rather than merely compressing frames. With a 1.4B-parameter generator trained only on publicly available academic motion--text data, \name outperforms all evaluated academic-data text-to-motion baselines and remains competitive with systems trained on much larger non-public motion corpora. Without task-specific retraining, the same formulation also improves prefix-conditioned generation, BABEL sequential rollout, and narrative motion composition. These results indicate that kinematic-unit latent factorization provides an effective generation substrate for controllable SMPL motion synthesis. Code will be released at https://github.com/ZeyuLing/PRISM.
Figures & tables
Fig. 2: Overview of the PRISM framework.
Method
HumanML3D [ 12 ]
Physical Plausibility
R-P T1 ↑
R-P T2 ↑
R-P T3 ↑
FID ↓
MM-D ↓
Div. →
Slide ↓
Float ↓
Jitter ↓
Real (SMPL)
0.770
0.903
0.944
0.0
14.88
27.77
2.00
2.6
5.5
HumanML3D-263-based methods
Real (HML3D → SMPL)
0.717
0.872
0.922
67.4
16.79
26.63
1.67
2.3
5.3
MotionGPT [ 16 ]
0.494
0.635
0.694
23.7
19.68
25.54
3.88
10.9
5.2
MotionGPT3 [ 51 ]
0.671
0.824
0.882
21.0
17.57
25.69
3.81
9.7
4.8
TABLE I: Text-to-motion on HumanML3D with the TMR evaluator.
Fig. 3: Qualitative comparison on text-to-motion. Red prompt words mark key action constraints, and red boxes/lines highlight visible mismatches or artifacts in the compared generations.
Prefix
Method
HumanML3D [ 12 ]
Physical Plausibility
R-P T1 ↑
R-P T2 ↑
R-P T3 ↑
FID ↓
MM D ↓
Div. →
Slide ↓
Float ↓
Jitter ↓
–
Real motion
0.770
0.903
0.944
0.0
14.88
27.77
2.00
2.6
5.5
1 frame
KIMODO [ 31 ]
0.525
0.690
0.769
113.0
21.06
25.97
2.20
3.4
5.4
FlowMDM [ 1 ]
0.449
0.630
0.706
89.9
19.87
25.77
3.72
8.0
6.2
MotionLab [ 13 ]
0.526
0.671
0.744
58.3
18.79
27.11
1.97
3.6
5.0
MotionStreamer [ 40 ]
0.617
0.780
0.850
14.7
16.93
27.00
5.67
18.4
11.5
TABLE II: Prefix-pose-conditioned generation on HumanML3D.
Method
Full BABEL
Long-Horizon Stress Slices
R@3 ↑
FID ↓
MM D ↓
T-FID ↓
Long R@3 ↑
Long FID ↓
Act. R@3 ↑
Act. FID ↓
Trans. FID ↓
Real
0.653
0.0
19.73
0.0
0.632
0.0
0.621
0.0
0.0
DoubleTake [ 33 ]
0.448
63.4
22.47
61.4
0.368
60.5
0.343
62.7
55.8
FlowMDM [ 1 ]
0.522
29.3
21.04
33.5
0.485
39.4
0.473
36.4
40.1
MotionStreamer [ 40 ]
0.397
51.3
22.21
68.0
0.321
64.1
0.308
59.2
75.1
PRISM (ours)
0.625
28.6
20.09
38.9
0.614
26.6
0.591
29.0
36.8
TABLE III: Sequential action generation on official BABEL validation episodes.
Method
Reconstruction (HumanML3D)
ICR
MPJPE ↓
RA-MPJPE ↓
dRoot-XZ ↓
dRoot-Y ↓
PA-MPJPE ↓
MPJRE ↓
rFID ↓
Emb-L2 ↓
MLD [ 6 , 9 ]
118.54
44.29
3.61
2.41
33.62
8.43
53.19
13.32
–
MoGenTS [ 43 ]
104.42
32.17
1.81
0.90
20.99
7.19
46.07
11.52
0.16×
MotionStreamer [ 40 ]
40.02
25.88
0.59
1.73
19.18
5.21
2.08
4.19
1.06×
PRISM (ours)
9.72
1.74
0.50
0.24
1.09
0.30
1.04
1.51
1.5×
TABLE IV: Paired motion reconstruction on HumanML3D.
Fig. 4: Qualitative comparison on long-horizon narrative composition. Colors denote consecutive atomic action segments from the shared segment plan, and red boxes highlight missed or unnatural actions in the baseline sequence.
Latent
Recon.
T2M Gen.
rFID ↓
MPJPE ↓
FID ↓
R-P T3 ↑
MM-D ↓
1D (monolithic)
8.90
45.13
36.7
0.822
17.08
2D (joint, ours)
1.04
9.72
10.1
0.925
15.51
TABLE V: Kinematic-unit-factorized versus monolithic latent spaces under controlled reconstruction and generation settings.
KAFS Mode
FID ↓
R-P T3 ↑
MM-D ↓
Div. →
None (baseline)
12.1
0.925
15.47
27.35
Uniform ( α=1.0 )
11.0
0.926
15.46
27.36
Random ( α∼U )
11.4
0.914
15.54
27.22
Depth-driven (ours)
10.1
0.925
15.51
27.44
TABLE VI: Kinematic-adaptive flow scheduling ablation on HumanML3D with the TMR evaluator.
Setting
HumanML3D [ 12 ]
BABEL Seq
R-P T1 ↑
R-P T2 ↑
R-P T3 ↑
FID ↓
MM-D ↓
Div. →
R@3 ↑
FID ↓
MM-D ↓
T-FID ↓
GT
0.770
0.903
0.944
0.0
14.88
27.77
0.653
0.0
19.73
0.0
2D RoPE
0.730
0.855
0.908
12.9
16.70
27.33
0.598
45.8
21.27
43.6
SMPL Index
0.739
0.861
0.914
12.4
16.08
26.94
0.605
37.3
20.64
40.9
KT-RoPE
0.752
0.878
0.925
10.1
15.51
27.44
0.625
28.6
20.09
38.9
TABLE VII: RoPE design comparison.
Criterion
vs. KIMODO
vs. MotionStreamer
W
T
L
W
T
L
Motion quality
43.7
36.1
20.2
78.4
11.7
9.9
Text fidelity
76.8
9.7
13.5
69.2
17.1
13.7
Transition
69.4
17.3
13.3
72.4
15.7
11.9
Overall
73.2
17.9
8.9
77.1
14.8
8.1
TABLE VIII: Narrative motion composition user study. W/T/L reports the percentage of pairwise win, tie, and loss votes for PRISM.
Text-to-motion generation requires modeling both global action structure and fine-grained motion dynamics from natural language. Existing approaches typically rely on either continuous diffusion models or vector-quantized discrete representations. Diffusion models generate smooth motions but lack explicit compositional structure for temporal planning, while discrete token-based methods improve controllability but compress motion into finite codebooks, losing fine-grained dynamics. We argue that this limitation stems from a representation mismatch: action semantics such as intent, phase transitions, and temporal layout are inherently discrete and compositional, whereas joint trajectories and motion dynamics are continuous and locally correlated. To address this, we propose DC-Motion, a discrete-continuous factorized framework for human motion generation. DC-Motion decomposes motion into discrete structural tokens capturing global action layout and continuous residual latents modeling fine-grained dynamics. A text-conditioned structure generator predicts discrete tokens via iterative masked modeling, and a diffusion-based residual generator produces continuous motion conditioned on the structure. Experiments on HumanML3D and KIT-ML demonstrate that DC-Motion achieves strong performance in both FID and R-Precision, outperforming representative diffusion-based and discrete-token baselines.
Hequan Wang, Xuean Chen, Jiaxu Zhang +2
State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430072, China · Geophysical Exploration Brigade of Hubei Provincial Geological Bureau, Wuhan 430058, China · Information Systems Technology and Design Pillar, Singapore University of Technology and Design, 487372, Singapore
Text-to-motion generation is driven by learning motion representations for semantic alignment with language. Existing methods rely on either continuous or discrete motion representations. However, continuous representations entangle semantics with dynamics, while discrete representations lose fine-grained motion details. In this context, we propose FlowCoMotion, a novel motion generation framework that unifies both treatments from a modeling perspective. Specifically, FlowCoMotion employs token-latent coupling to capture both semantic content and high-fidelity motion details. In the latent branch, we apply multi-view distillation to regularize the continuous latent space, while in the token branch we use discrete temporal resolution quantization to extract high-level semantic cues. The motion latent is then obtained by combining the representations from the two branches through a token-latent coupling network. Subsequently, a velocity field is predicted based on the textual conditions. An ODE solver integrates this velocity field from a simple prior, thereby guiding the sample to the potential state of the target motion. Extensive experiments show that FlowCoMotion achieves competitive performance on text-to-motion benchmarks, including HumanML3D and SnapMoGen.
Dawei Guan, Di Yang, Chengjie Jin +1
School of Artificial Intelligence and Data Science, University of Science and Technology of China · Suzhou Institute for Advanced Research, University of Science and Technology of China
We introduce MoLingo, a text-to-motion (T2M) model that generates realistic, lifelike human motion by denoising in a continuous latent space. Recent works perform latent space diffusion, either on the whole latent at once or auto-regressively over multiple latents. In this paper, we study how to make diffusion on continuous motion latents work best. We focus on two questions: (1) how to build a semantically aligned latent space so diffusion becomes more effective, and (2) how to best inject text conditioning so the motion follows the description closely. We propose a semantic-aligned motion encoder trained with frame-level text labels so that latents with similar text meaning stay close, which makes the latent space more diffusion-friendly. We also compare single-token conditioning with a multi-token cross-attention scheme and find that cross-attention gives better motion realism and text-motion alignment. With semantically aligned latents, auto-regressive generation, and cross-attention text conditioning, our model sets a new state of the art in human motion generation on standard metrics and in a user study. We will release our code and models for further research and downstream usage.
Yannan He, Garvita Tiwari, Xiaohan Zhang +4
University of Tübingen, Germany · Tübingen AI Center, Germany · Max Planck Institute for Informatics, Saarland Informatics Campus, Germany +2