TopX: Text-to-Motion for Unseen Skeletons from the Skeleton File Alone

Tengjiao Sun; Pengcheng Fang; Xiaoyu Zhan; Dongjie Fu; Xiaohao Cai; Hansung Kim

cs.CVOct 25, 2026

TopX: Text-to-Motion for Unseen Skeletons from the Skeleton File Alone

Authors: Tengjiao Sun, Pengcheng Fang, Xiaoyu Zhan, Dongjie Fu, Xiaohao Cai, Hansung Kim

Organizations: University of Southampton · MOGO AI · Nanjing University

Abstract

Text-to-motion models generate motion only for the skeletons they were trained on, while a new character arrives as a skeleton file and nothing else. We study text-to-motion for skeletons unseen in training, without clips of the character, templates, retargeting or fine-tuning. We present TopX, a 73M-parameter flow-matching transformer that generates joint rotations and a root trajectory for any skeleton from its kinematic tree, rest pose and joint names. Motion is normalised by the rest pose and the size of the skeleton, and the rest pose is encoded as a world-frame descriptor that is independent of how the rig's author oriented each joint, while rest conventions are randomised during training. Attention is factorised over the joints and the frames of a motion. Across joints, queries and keys are rotated by HK-RoPE, a heat-kernel rotary encoding of the skeleton graph that is defined for every tree and does not change with the order of its joints or the eigenbasis of its Laplacian; across frames, they are rotated by a rotary encoding of frame order. To evaluate this generalisation, we build SkelShift, which unifies four motion corpora under one codec and holds out whole skeletons, tree classes and species, so that each test bucket isolates one kind of skeleton shift: topology, joint count, body plan, species or rig source. It scores motion quality without a learned evaluator. A single model animates every benchmark skeleton (4--117 joints) directly from its skeleton file and generates up to 240 frames. On held-out game species its motion is about as close to the real clips as two halves of the real clips are to each other (kernel distance 0.127 against 0.124), and text-to-motion retrieval reaches R@1 0.953, compared with 0.929 for real motion. Project page: https://chdtevior.github.io/TopX-page/.

Figures & tables

Figure 1: TopX. ❶ No clips of an unseen skeleton are needed. Its tree, rest pose and joint names give each joint a feature $e_{j}$ (an MLP of tree and rest-pose features plus the LLM2Vec embedding of the joint’s description). The motion tokens, one per joint and frame (position, rotation relative to the rest pose, velocity and contact; ground track and heading on the root), are normalised by the rest pose and the skeleton’s size ( Eq. 1 ), so that the rest frame maps to zero. ❷ Each of ten blocks attends over time with queries and keys rotated by $\theta_{f}=f\omega$ , and over joints with queries and keys rotated by HK-RoPE, whose angle $\theta_{j}$ is an MLP of joint $j$ ’s heat-kernel signature at eight scales on the tree, together with a bias on tree hops. Mesh colours show the first principal component of the log signature, on a colour scale shared by all skeletons. The caption and the flow time $t$ condition the network through AdaLN. Sampling from noise yields the motion shown on Stego, a Truebones test skeleton. Meshes are for display only; the human mesh is SMPL+H [ 36 ] .

Bucket	Source	Unseen at test time	Clips dev / test	Rigs	Trees	Joints	TED	Reused motion
ID-clip	UniML3D	clips of training rigs	70 / 172	135	79	10–60	0	26%
ID-tree	UniML3D	rigs whose tree is seen	255 / 508	759	140	6–60	0	63%
OOD-tree	UniML3D	tree classes (bipeds, quadrupeds)	213 / 320	333	266	11–60	2	15%
OOD-plan	UniML3D	tree classes of rare body plans	40 / 62	70	63	5–60	5	2%
OOD-J $>$ 60 (Mixamo)	UniML3D	rigs of one 65-joint humanoid tree	155 / 350	250	1	65	5	51%
OOD-J $>$ 60 (other)	UniML3D	rigs above the corpus’s joint range	166 / 512	464	184	61–117	7	33%

Table 1: SkelShift buckets. A bucket holds out units —clips, rigs, classes of isomorphic trees or species—and every clip of a held-out unit is removed from training and validation. TED: median tree edit distance from a clip’s tree to the nearest training tree. Reused motion: share of clips whose source animation appears in training (detected by motion fingerprints for UniML3D and by matching joint rotations for the game species).

			UniMate tuned on FID kin						UniMate tuned on FID desc
			FID desc $\downarrow$		FID kin $\downarrow$		Pen. $\downarrow$		FID desc $\downarrow$		FID kin $\downarrow$		Pen. $\downarrow$
Test clips	Clips	R–R	Ours	Uni.	Ours	Uni.	Ours	Uni.	Ours	Uni.	Ours	Uni.	Ours	Uni.
All	1,062	0.48	4.83	6.06	2.00	2.05	0.24	0.33	4.83	4.32	2.00	2.52	0.24	0.28
ID-clip	172	2.83	7.18	9.09	2.40	2.49	0.25	0.34	7.18	5.96	2.40	2.57	0.25	0.27
No training clip	890	0.55	4.58	5.73	0.55	0.48	0.24	0.33	4.58	4.22	0.55	1.48	0.24	0.28
novel motion	453	1.13	9.63	10.70	1.15	0.89	0.27	0.35	9.63	7.95	1.15	2.17	0.27	0.29

Table 2: Comparison of TopX and UniMate, both trained on the same 5,023 UniML3D clips (test clips of the four UniML3D buckets, first 60 frames, five seeds; additional metrics: appendix), with UniMate’s guidance tuned on dev by FID kin (left) or by FID desc (right). Bold: the lower value within each half, where the paired 95% confidence interval over rigs excludes zero. R–R: FID desc between two halves of the real clips. Pen.: share of frames below the ground (0 for real clips).

Figure 2: Samples on unseen test skeletons, each drawn on the rig’s own mesh: five frames at equal times, light to dark. (a) The two models of Tab. 2 (UniMate tuned on FID kin ) on four rigs that neither was trained on, TopX in the upper row, over at most 60 frames. (b) The main model on two held-out game species, a capybara and an emu on their production meshes; the camera follows the animal, and squares show its root path from above. Rendering details, further examples and mesh credits are given in Appendix J .

Bucket	Rigs / clips	KID desc $\downarrow$ (real–real)	Slip $\rightarrow$ (real)	Pen. $\downarrow$	Float $\rightarrow$ (real)	Speed $\rightarrow$
ID-clip	93 / 172	0.950 (0.455)	0.309 (0.353)	0.204	0.110 (0.128)	0.87
ID-tree	504 / 508	0.145 (0.002)	0.304 (0.349)	0.172	0.021 (0.039)	0.90
OOD-tree	218 / 320	1.74 (0.142)	0.310 (0.366)	0.200	0.045 (0.078)	0.77
OOD-plan	45 / 62	10.9 (6.04)	0.176 (0.280)	0.251	0.106 (0.164)	0.81
OOD-J $>$ 60 (Mixamo)	172 / 350	0.664 (0.874)	0.311 (0.363)	0.131	0.024 (0.050)	0.82
OOD-J $>$ 60 (other)	351 / 512	0.240 (0.092)	0.275 (0.333)	0.162	0.021 (0.029)	0.77

Table 3: Results of TopX on every test bucket at full length (up to 240 frames, five seeds per clip). Real-clip values are in parentheses (

\rightarrow

: closer to the real value is better), and KID desc is shown with its real–real value. ∗ Fewer than 20 rigs. The single OOD-J

<

5 test clip is included only in the last row.

Arm	OOD- tree	OOD- plan	New species	OOD- rig	Rest $\downarrow$
Real	0.142	6.04	104	0.914	–
Full model	1.78	11.1	82.4	1.01	0.44
Learned frame table	1.88	11.7	84.4	1.59	–
Learned joint table	1.87	11.5	83.7	1.12	–
Eigenvectors + SignNet	1.83	11.5	83.6	1.17	–
$-$ augmentation	1.95	12.1	91.8	1.03	0.89

Table 4: Ablations, each trained for half the steps of the main model and evaluated on the OOD buckets targeted by the changed components: KID desc on unseen topologies (OOD-tree), rare body plans (OOD-plan), new species and Truebones rigs (OOD-rig), five seeds per clip, up to 240 frames. Rest: median change of the output, in mean bone lengths, when a test rig’s rest convention is rotated ( Appendix I ). Real: KID desc between two halves of the real clips. Confidence half-widths are given in Appendix I .

Appendix figures & tables14 assets

Supplementary material from the paper’s appendix.

Appendix

Slice	Content	Rows
0:3	position: XZ from the smoothed root track, Y absolute	all
3:9	rotation $\mathbf{R}_{j}\mathbf{R}^{\mathrm{rest}\top}_{j}$ , continuous 6D	all
9:12	world velocity (supervision only)	all
12	contact: height $\leq 0.05\,s$ , speed $\leq 0.25\,s$ /s	all
13:15	smoothed root track XZ	root
15:17	heading $(\cos\phi,\sin\phi)$	root

Appendix

Table 5: Channels per joint and frame. Channels 13–16 are used on the root row only. Positions, velocities and the ground track are in the skeleton’s units;

s

is the diagonal of the rest pose’s bounding box.

Width $D$ , blocks $N_{b}$ , heads	512, 10, 8 (73.3M parameters)
Frames	1 rest frame + $L{=}240$ target frames
Heat-kernel diffusion times	$\tau_{i}=2^{i-2}$ , $i=0,\dots,7$
Temporal rotary base	700
Caption dropout	0.1
Epochs, draws per epoch	600, 55,984 (437 steps of 128; the last 48 dropped)
Optimiser	AdamW, weight decay 0.01

Appendix

Table 6: Settings of TopX.

Bucket	T0	T1	T2a	T2b
ID-clip, ID-tree	1,005	–	–	–
OOD-tree	–	169	145	219
OOD-plan	–	8	32	62
OOD-J $>$ 60 (Mixamo)	–	505	–	–
OOD-J $>$ 60 (other)	–	356	120	202
OOD-J $<$ 5	–	1	–	1

Appendix

Table 7: Topology tiers per bucket (number of clips, dev and test combined).

	FID desc $\downarrow$	Skate $\rightarrow$	Slide $\rightarrow$	Cov. $\rightarrow$	Pen. $\downarrow$	Float $\rightarrow$	Speed $\rightarrow$	KID desc $\downarrow$
Real clips	–	0.254	1.71	0.442	0.000	0.050	1	0.091
$w{=}1.5$	1.88	0.231	1.60	0.452	0.179	0.025	0.78	0.149
$w{=}2$	1.71	0.243	1.74	0.441	0.170	0.030	0.84	0.142
$w{=}3$	1.79	0.260	2.02	0.417	0.172	0.039	0.95	0.176

Appendix

Table 8: TopX on all dev buckets (638 rigs, 1,563 clips, full length, five seeds per clip, 50 Euler steps) at three guidance scales. The real-clip row gives the real values and, under KID desc , the real–real value.

	Evaluations	Seconds
TopX, 20 steps	40	0.67
TopX, 50 steps	100	1.63
TopX, 100 steps	200	3.23
TopX, 200 steps	400	6.43
TopX, 400 steps	800	12.82
UniMate, guidance 3	366.4	6.13

Appendix

Table 9: Network evaluations and seconds per sample on one H200 NVL (50 dev clips, seed 7, batch 1; TopX: the model of Tab. 2 , at the clip’s length). Under classifier-free guidance, each Euler step and each function evaluation of UniMate’s solver requires two network evaluations.

		FID desc			FID kin , Motion2Motion		FID kin , Bailando
Test clips	Clips	Ours	UniMate	R–R	Ours	UniMate	Ours	UniMate
All	1,062	4.83	6.06	0.48	2.00	2.05	1.85	1.90
ID-clip	172	7.18	9.09	2.83	2.40	2.49	2.10	2.14
No training clip	890	4.58	5.73	0.55	0.55	0.48	0.56	0.44
novel motion	453	9.63	10.70	1.13	1.15	0.89	1.05	0.77
reused motion	437	0.65	1.64	0.60	0.06	0.24	0.08	0.27

Appendix

Table 10: Same-data comparison, Fréchet distances (test, first 60 frames, five seeds; mean over seeds). Ours: 200 Euler steps at guidance 2, at the clip’s length; UniMate: guidance 7. FID kin on the Motion2Motion and the Bailando release of the kinetic features. Bold: the lower value, where the paired 95% confidence interval over rigs excludes zero. R–R: real clips, half against half (groups of at least 100 clips).

	Pen. $\downarrow$		Skate			Cov.			Speed
Test clips	Ours	UniMate	Ours	UniMate	Real	Ours	UniMate	Real	Ours	UniMate
All	0.24	0.33	0.33	0.35	0.35	0.60	0.58	0.64	0.93	1.08
ID-clip	0.25	0.34	0.34	0.37	0.33	0.52	0.51	0.55	0.98	1.23
No training clip	0.24	0.33	0.33	0.34	0.35	0.61	0.59	0.66	0.92	1.06
novel motion	0.27	0.35	0.26	0.33	0.30	0.52	0.47	0.57	0.90	1.13
reused motion	0.23	0.32	0.39	0.35	0.39	0.69	0.70	0.75	0.94	1.02

Appendix

Table 11: Same-data comparison, ground contact and activity (test, first 60 frames, five seeds; Ours and UniMate as in Tab. 10 ). Pen.: share of frames below the ground (0 for real clips). Skate and Cov.: closer to the real value is better. Speed: joint speed relative to the real clips. Bold: lower Pen., or a smaller absolute difference from the real value for Skate and Cov., where the paired 95% confidence interval over rigs excludes zero.

		FID desc $\downarrow$		FID kin M2M $\downarrow$		FID kin Bk. $\downarrow$		Pen. $\downarrow$		Skate			Speed
Test clips	Clips	Ours	Uni.	Ours	Uni.	Ours	Uni.	Ours	Uni.	Ours	Uni.	Real	Ours	Uni.
All	1,062	4.83	4.32	2.00	2.52	1.85	2.18	0.24	0.28	0.33	0.24	0.35	0.93	0.85
ID-clip	172	7.18	5.96	2.40	2.57	2.10	2.22	0.25	0.27	0.34	0.25	0.33	0.98	0.93
No training clip	890	4.58	4.22	0.55	1.48	0.56	1.13	0.24	0.28	0.33	0.24	0.35	0.92	0.83
novel motion	453	9.63	7.95	1.15	2.17	1.05	1.70	0.27	0.29	0.26	0.21	0.30	0.90	0.83
reused motion	437	0.65	1.21	0.06	0.72	0.08	0.47	0.23	0.27	0.39	0.27	0.39	0.94	0.83

Appendix

Table 12: Same-data comparison with UniMate tuned on dev FID desc (guidance 3; test, first 60 frames, five seeds; Ours as in Tab. 10 ). FID kin : Motion2Motion (M2M) and Bailando (Bk.) releases of the kinetic features. Skate: closer to the real value is better; Speed: relative to the real clips (target 1). Bold: lower FID desc , FID kin or Pen., or closer to the real value for Skate and Speed, where the paired 95% confidence interval over rigs excludes zero.

Method	FID desc $\downarrow$	Skate $\rightarrow$	Slide $\rightarrow$	Cov. $\rightarrow$
Rigs unseen by all three: capybara and emu, 6 rigs, 635 clips
Real clips	–	0.213	1.03	0.199
AnyTop † [ 12 ]	24.63	0.759	10.65	0.358
UniMate [ 31 ]	3.32	0.268	1.59	0.220
TopX	1.19	0.208	1.27	0.234
Trained on by both baselines: 6 Truebones rigs, 62 clips

Appendix

Table 13: Comparison with released weights (test, first 60 frames, five seeds per clip). TopX is the main model, trained on four corpora. Top: the two held-out species, on rigs that none of the three methods was trained on; bottom: the Truebones test rigs, on which both baselines were trained. † Unconditional and transductive. Real clips give the reference values.

Bucket	Rigs / clips	Real–real	Full	Frame table	Joint table	Eigvec. + SignNet	$-$ aug.
OOD-tree	218 / 320	0.142	1.78 $\pm$ 0.03	1.88 $\pm$ 0.03	1.87 $\pm$ 0.04	1.83 $\pm$ 0.03	1.95 $\pm$ 0.04
OOD-plan	45 / 62	6.04	11.1 $\pm$ 0.2	11.7 $\pm$ 0.1	11.5 $\pm$ 0.3	11.5 $\pm$ 0.2	12.1 $\pm$ 0.2
OOD-new-species ∗	9 / 47	104	82.4 $\pm$ 0.8	84.4 $\pm$ 1.0	83.7 $\pm$ 0.7	83.6 $\pm$ 1.3	91.8 $\pm$ 1.8
OOD-rig ∗	6 / 62	0.914	1.01 $\pm$ 0.11	1.59 $\pm$ 0.04	1.12 $\pm$ 0.11	1.17 $\pm$ 0.15	1.03 $\pm$ 0.03

Appendix

Table 14: Table 4 with its uncertainty: KID desc as the mean over the five sampling seeds

\pm

its 95% half-width over them. Columns: full model, learned frame table, learned joint table, eigenvectors + SignNet, without augmentation. ∗ Fewer than 20 rigs.

Figure 3: One test clip from each of five OOD buckets, showing the real clip and the output of TopX (guidance 2, seed 7). Each strip shows the real clip above ours at matching times; the strobe panel (Mixamo) shows the real clip on the left and overlays five poses spaced along the real clip’s root path. Headers give the bucket and the caption.

Figure 4: Four additional humanoid rigs without training clips, drawn as in Fig. 2 (a), with TopX above UniMate.

Mesh (author)	Licence	Figure
Defeat Idle (vicente betoret ferrero)	CC BY 4.0	2 (a)
Lowpoly Dog Walk (timesen)	CC BY 4.0	2 (a)
Dey Flair Dancing (jdey123)	CC BY 4.0	2 (a)
Back Flip To Uppercut (1) (marcolinadesign)	CC BY 4.0	2 (a)
Low poly fox running animation (dragonsnap)	CC BY 4.0	3
fario3 (masterdef)	CC BY 4.0	3

Appendix

Table 15: Meshes under public licences in Figs. 1 , 2 , 3 and 4 ; the human mesh in Fig. 1 is SMPL+H [ 36 ] . † Title shortened.

	All		Emu
	R@1	Shuf.	R@1	Shuf.
Real clips, first 60 frames	0.896	–	0.871	–
Real clips, full length	0.929	–	0.933	–
TopX, generating 60 frames	0.915 $\pm$ 0.007	0.122	0.928 $\pm$ 0.006	0.137
TopX, clip length, first 60 frames	0.932 $\pm$ 0.007	0.113	0.922 $\pm$ 0.016	0.127
TopX, clip length, full	0.953 $\pm$ 0.004	0.120	0.954 $\pm$ 0.013	0.137

Appendix

Table 16: Text-to-motion R@1 within rigs of the held-out game species (pools of 8, chance 0.125): mean and 95% half-width over five seeds, and the R@1 with captions shuffled (Shuf.). All: 439 clips on 6 rigs; emu: 181 clips on 3 rigs.

Text-to-Motion Generation

Text-to-motion generation converts natural-language descriptions into semantically aligned, temporally coherent 3D human movement.

See all 96 papers

UniMate: One Unified Model to Animate Diverse Skeletons
Sep 4, 2026Linzhan Mou, Jiahui Lei, Zhiyang Dou +4Text-to-Motion Generation Motion Diffusion Models
Princeton University, USA · University of California, Berkeley, USA · Massachusetts Institute of Technology, USA +1
TAME:Topology-Aware Text-Driven Motion Editing across Heterogeneous Humanoid Skeletons
Oct 3, 2026Qichen Zheng, Siyuan Yang, Chong Wang +4Human Motion Generation
Nanyang Technological University · KTH Royal Institute of Technology · Lancaster University
SAMoR: Motion Modelling for Articulated Objects of Any Skeleton and Topology
Jul 2, 2026Yuhao Zhang, Gerard Pons-Moll, Tolga BirdalGenerative Modeling Motion Representation Learning
Imperial College London · University of Tübingen, Tübingen AI Center

Sep 4, 2026J/K move · Enter open · S save

UniMate: One Unified Model to Animate Diverse Skeletons

Linzhan Mou, Jiahui Lei, Zhiyang Dou, Chenyue Cai+3

Princeton University, USA · University of California, Berkeley, USA · Massachusetts Institute of Technology, USA+1

Text-to-Motion Generation Motion Diffusion Models Diffusion Transformer Human Motion Generation

Open paper AskPDF arXiv

Recent advances in automatic rigging now deliver animation-ready 3D assets at scale, yet generating the motion to drive them remains a bottleneck. Existing learned animators are topology-constrained: they rely on category-specific templates or require per-skeleton fine-tuning and reference motions at inference. We present UniMate, a unified foundation model that synthesizes articulated motion for arbitrary skeletons from a rigged 3D asset and a text prompt, with no test-time optimization or per-skeleton retraining. UniMate introduces a topology-aware diffusion transformer, which integrates skeletal topology into attention via three mechanisms: (1) a graph-aware attention bias from pairwise joint relations and geodesic distances; (2) a spectral rotary position embedding generalizing RoPE to arbitrary kinematic trees via the graph Laplacian; and (3) a global topological conditioner attention-pooled from the rest-pose skeleton. We also curate UniML3D, 13,006 motion sequences spanning bipedal, quadrupedal, avian, marine, insectoid, serpentine, and articulated rigid objects with unified canonicalization and text pairing. Trained on this dataset, UniMate outperforms state-of-the-art baselines in quality, generalization, and efficiency, and supports zero-shot cross-topology transfer, in-betweening, expansion, and text-guided editing. Our project page is available at https://linzhanmou.com/unimate/.

Abstract

Text-to-Motion Generation

Explore similar work

UniMate: One Unified Model to Animate Diverse Skeletons

TAME:Topology-Aware Text-Driven Motion Editing across Heterogeneous Humanoid Skeletons

SAMoR: Motion Modelling for Articulated Objects of Any Skeleton and Topology