Text-to-motion mapping on HumanML3D (test)
0.002FIDReal motions
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Real motionsMotion Length=Real2022.08 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| RealCondition=Ground Truth2022.09 | 0.002 | — | — | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motionsMotion length=Ground Truth2023.04 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real2023.06 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motion2023.09 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real2023.09 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motion2023.01 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motions2023.04 | 0.002 | — | — | — | 2.974 | 9.503 | — | 51.1 | — | |
| Real2023.09 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motion2023.11 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motions2023.12 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motion2024.01 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motion2024.03 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real MotionCategory=Reference2026.02 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motions2022.07 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real motionGround truth motion length=True2024.06 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | — | — | — | — | |
| Real motion2026.03 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| Real2026.03 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| RealCategory=N/A, #param=-2026.06 | 0.002 | 51.1 | 70.3 | 79.7 | 2.974 | 9.503 | — | — | — | |
| LaMPCategory=Separated Model2026.02 | 0.032 | 55.7 | 75.1 | 84.3 | 2.759 | 9.571 | — | — | — | |
| LAMPMethod Category=Mask-Modeling-based Motion Generator, w/ GRPO=false2025.12 | 0.032 | 55.7 | 75.1 | 84.3 | 2.759 | 9.571 | — | — | — | |
| MoGenTSCategory=T2M Only2026.02 | 0.033 | 52.9 | 71.9 | 81.2 | 2.867 | 9.57 | — | — | — | |
| MoGenSTMethod Category=Mask-Modeling-based Motion Generator, w/ GRPO=false2025.12 | 0.033 | 52.9 | 71.9 | 81.2 | 2.867 | 9.57 | — | — | — | |
| ParTY2026.03 | 0.035 | 55 | 74.4 | 83.6 | 2.779 | 9.534 | 2.155 | — | — | |
| Light-T2MFramework=Volume Diffusion, Ground truth motion length=True2024.06 | 0.04 | 51.1 | 69.9 | 79.5 | 3.002 | — | 1.67 | — | — | |
| MoMaskbase-layer tokens only=false2023.11 | 0.045 | 52.1 | 71.3 | 80.7 | 2.958 | — | 1.241 | — | — | |
| MoMaskCategory=T2M Only2026.02 | 0.045 | 52.1 | 71.3 | 80.7 | 2.958 | 9.62 | — | — | 1.241 | |
| MoMask2026.02 | 0.045 | 52.1 | 71.3 | 80.7 | 2.958 | — | 12.41 | — | — | |
| MoMaskMethod Category=Mask-Modeling-based Motion Generator, w/ GRPO=false2025.12 | 0.045 | 52.1 | 71.3 | 80.7 | 2.958 | — | — | — | — | |
| MoMaskFramework=VAE, Ground truth motion length=True2024.06 | 0.045 | 52.1 | 71.3 | 80.7 | 2.958 | — | 1.241 | — | — | |
| MoMask2026.03 | 0.045 | 52.1 | 71.3 | 80.7 | 2.958 | — | 1.241 | — | — | |
| MoMaskCategory=Text-only, #param=45M2026.06 | 0.045 | 52.1 | 71.3 | 80.7 | 2.958 | — | — | — | — | |
| DiMo w/ GRPOCategory=Unified Model, GRPO=true2026.02 | 0.047 | 52.8 | 72.4 | 81.8 | 2.862 | 9.419 | — | — | 2 | |
| DiMo w/o GRPOCategory=Unified Model, GRPO=false2026.02 | 0.05 | 52.8 | 72.3 | 81.8 | 2.867 | 9.515 | — | — | 2.016 | |
| MCMCategory=Multimodal, #param=-2026.06 | 0.053 | 50.2 | 69.2 | 78.8 | 3.037 | 9.585 | — | — | — | |
| BAMMMethod Category=Mask-Modeling-based Motion Generator, w/ GRPO=false2025.12 | 0.055 | 52.5 | 72 | 81.4 | 2.919 | 9.717 | — | — | — | |
| BAMM2026.03 | 0.055 | 52.5 | 72 | 81.4 | 2.919 | 9.717 | 1.687 | — | — | |
| MARDMMethod Category=Diffusion-based Motion Generator, w/ GRPO=false2025.12 | 0.061 | 52.3 | 71.5 | 81 | — | — | — | — | — | |
| VQ-VAE2023.09 | 0.064 | 49.1 | 68.4 | 78.1 | 3.097 | 9.653 | — | — | — | |
| T2M Mamba2026.02 | 0.068 | 50.6 | 69.6 | 79.3 | 3.034 | 9.497 | 23.1 | — | — | |
| VQ-VAEtraining without WMC=true2023.09 | 0.07 | 49.6 | 68.9 | 78.7 | 3.072 | 9.593 | — | — | — | |
| Oursclassifier-free scale=12023.09 | 0.07 | 49.6 | 68.7 | 78.3 | 3.063 | 9.551 | 2.062 | — | — | |
| Our VQ-VAE (Recons.)2023.01 | 0.07 | 50.1 | 69.2 | 78.5 | 3.072 | 9.593 | — | — | — | |
| Light-RDM-1Framework=Volume Diffusion, Ground truth motion length=True2024.06 | 0.07 | 51 | 69.9 | 79.4 | 3.005 | — | 1.713 | — | — | |
| Oursclassifier-free scale=22023.09 | 0.072 | 51.5 | 70.6 | 80.2 | 2.941 | 9.683 | 1.869 | — | — | |
| MoTeCategory=Unified Model2026.02 | 0.075 | 54.8 | 73.7 | 82.5 | 2.867 | — | — | — | 2.399 | |
| MMM2026.03 | 0.08 | 50.4 | 69.6 | 79.4 | 2.998 | 9.411 | 1.164 | — | — | |
| MMMCategory=Text-only, #param=254M2026.06 | 0.08 | 50.4 | 69.6 | 79.4 | 2.998 | 9.411 | — | — | — | |
| MoMaskbase-layer tokens only=true2023.11 | 0.082 | 50.4 | 69.9 | 79.7 | 3.05 | — | 1.05 | — | — | |
| MaskControlCategory=T2M Only2026.02 | 0.083 | — | — | 80.5 | — | 9.395 | — | — | — | |
| MMMMethod Category=Mask-Modeling-based Motion Generator, w/ GRPO=false2025.12 | 0.089 | 51.5 | 70.8 | 80.4 | 2.926 | 9.577 | — | — | — | |
| ReMoMaskCategory=T2M Only2026.02 | 0.099 | 53.1 | 72.2 | 81.3 | 2.865 | 9.535 | — | — | 2.823 | |
| ReMoDiffuseMotion length=Ground Truth2023.04 | 0.103 | 51 | 69.8 | 79.5 | 2.974 | 9.018 | 1.795 | — | — | |
| ReMoDiffuse2023.11 | 0.103 | 51 | 69.8 | 79.5 | 2.974 | — | 1.795 | — | — | |
| ReMoDiffuse2023.12 | 0.103 | 51 | 69.8 | 79.5 | 2.974 | 9.018 | 1.795 | — | — | |
| ReMoDiffuseMotion length source=ground-truth2024.03 | 0.103 | 51 | 69.8 | 79.5 | 2.974 | 9.018 | 1.795 | — | — | |
| ReMoDiffuseCategory=T2M Only2026.02 | 0.103 | 51 | 69.8 | 79.5 | 2.974 | 9.018 | — | — | 1.795 | |
| ReMoDiffuseMethod Category=Diffusion-based Motion Generator, w/ GRPO=false2025.12 | 0.103 | 51 | 69.8 | 79.5 | 2.974 | 9.018 | — | — | — | |
| ReMoDiffuseFramework=Volume Diffusion, Ground truth motion length=True2024.06 | 0.103 | 51 | 69.8 | 79.5 | 2.974 | — | 1.795 | — | — | |
| OdorikoCategory=Multimodal, #param=44M2026.06 | 0.103 | 51.2 | 70.9 | 80.5 | 2.93 | 9.672 | — | — | — | |
| GUESS2024.01 | 0.109 | 50.3 | 68.8 | 78.7 | 3.006 | 9.826 | 2.43 | — | — | |
| ParCo2024.03 | 0.109 | 51.5 | 70.6 | 80.1 | 2.927 | 9.576 | 1.382 | — | — | |
| ParCo2026.03 | 0.109 | 51.5 | 70.6 | 80.1 | 2.927 | 9.576 | 1.382 | — | — | |
| Ours (S2, w/ IRMoGen)Method Category=Motion-aware LLM, w/ GRPO=false, support both text and motion outputs=true2025.12 | 0.111 | 52.6 | 71.7 | 81 | 2.885 | 9.819 | — | — | — | |
| AttT2M2023.09 | 0.112 | 49.9 | 69 | 78.6 | 3.038 | 9.7 | 2.452 | — | — | |
| AttT2M2024.03 | 0.112 | 49.9 | 69 | 78.6 | 3.038 | 9.7 | 2.452 | — | — | |
| AttT2M2026.02 | 0.112 | 49.9 | 69 | 78.6 | 3.038 | 9.7 | 24.52 | — | — | |
| MARDM-SiTModel Category=Diffusion-based, Evaluator=MARDM [49]2025.12 | 0.114 | 50 | 69.5 | 79.5 | 3.27 | — | — | — | — | |
| MoLACategory=Text-only, #param=49M2026.06 | 0.115 | 51.6 | 71.2 | 80.5 | 3.008 | 9.885 | — | — | — | |
| T2M-GPTMotion length=Ground Truth2023.04 | 0.116 | 49.1 | 68 | 77.5 | 3.118 | 9.761 | 1.856 | — | — | |
| T2M-GPT2023.06 | 0.116 | 49.1 | 68 | 77.5 | 3.118 | 9.761 | 1.856 | — | — | |
| T2M-GPT2023.09 | 0.116 | 49.1 | 68 | 77.5 | 3.118 | 9.761 | 1.856 | — | — | |
| T2M-GPT2023.09 | 0.116 | 49.1 | 68 | 77.5 | 3.118 | 9.761 | 1.856 | — | — | |
| T2M-GPTtau value (τ)=0.52023.01 | 0.116 | 49.1 | 68 | 77.5 | 3.118 | 9.761 | 1.856 | — | — | |
| T2M-GPT2023.09 | 0.116 | 49.1 | 68 | 77.5 | 3.118 | 9.761 | 1.856 | — | — | |
| GraphMotiontotal diffusion steps=1502023.11 | 0.116 | 50.4 | 69.9 | 78.5 | 3.07 | 9.692 | 2.766 | — | — | |
| T2M-GPT2023.12 | 0.116 | 49.1 | 68 | 77.5 | 3.118 | 9.761 | 1.856 | — | — | |
| MoMaskModel Category=Mask-Modeling-based, Evaluator=MARDM [49]2025.12 | 0.116 | 49 | 68.7 | 78.6 | 3.353 | — | — | — | — | |
| T2M-GPT2026.03 | 0.116 | 49.1 | 68 | 77.5 | 3.118 | 9.761 | 1.856 | — | — | |
| GraphMotiontotal diffusion steps=502023.11 | 0.118 | 49.6 | 68.6 | 77.8 | 3.143 | 9.796 | 2.603 | — | — | |
| ViMoRAGMethod Category=Motion-aware LLM, w/ GRPO=false2025.12 | 0.131 | 45.2 | 65.5 | 76.4 | 3.146 | 9.424 | — | — | — | |
| MMMModel Category=Mask-Modeling-based, Evaluator=MARDM [49]2025.12 | 0.132 | 48.7 | 68.3 | 78.2 | 3.359 | — | — | — | — | |
| Ours (S1, w/ IRMoGen)Method Category=Motion-aware LLM, w/ GRPO=false, support both text and motion outputs=true2025.12 | 0.135 | 52.2 | 71.1 | 80.5 | 2.906 | 9.74 | — | — | — | |
| T2M-GPTtau value (τ)=02023.01 | 0.14 | 41.7 | 58.9 | 68.5 | 3.73 | 9.844 | 3.285 | — | — | |
| T2M-GPTtau value (τ)=U[0, 1]2023.01 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | 9.722 | 1.831 | — | — | |
| T2M-GPT2023.11 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | — | 1.831 | — | — | |
| T2M-GPT2023.11 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | 9.722 | 1.831 | — | — | |
| T2M-GPT2024.01 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | 9.722 | 1.831 | — | — | |
| T2M-GPT2024.03 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | 9.722 | 1.831 | — | — | |
| T2M-GPTCategory=T2M Only2026.02 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | 9.722 | — | — | 1.831 | |
| T2M-GPT2026.02 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | 9.722 | 18.31 | — | — | |
| Being-M0.5Method Category=Motion-aware LLM, w/ GRPO=false, support both text and motion outputs=true2025.12 | 0.141 | 52.8 | — | 81.5 | 2.953 | — | — | — | — | |
| T2M-GPTMethod Category=Motion-aware LLM, w/ GRPO=false2025.12 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | 9.722 | — | — | — | |
| T2M-GPTCategory=Text-only, #param=248M2026.06 | 0.141 | 49.2 | 67.9 | 77.5 | 3.121 | 9.722 | — | — | — | |
| Phase ControlNet (Flow)AITS=0.1572026.03 | 0.146 | 53.1 | 71.9 | 80.8 | 2.88 | 9.318 | — | — | — | |
| FineMoGen2023.12 | 0.151 | 50.4 | 69 | 78.4 | 2.998 | 9.263 | 2.696 | — | — | |
| MotionGPTEvaluation Protocol=Pre-trained2023.06 | 0.16 | 43.5 | 60.7 | 70 | 3.7 | 9.411 | 3.437 | — | — | |
| Ours (S1, w/ IRMoGen)Model Category=LLM-based, Evaluator=MARDM [49], Support both text and motion outputs=true2025.12 | 0.161 | 50.7 | 69.6 | 79.1 | 3.234 | — | — | — | — | |
| Being-M0Method Category=Motion-aware LLM, w/ GRPO=false2025.12 | 0.166 | 51.9 | — | 80.3 | 2.964 | — | — | — | — | |
| MotionCraftCategory=Multimodal, #param=77M2026.06 | 0.173 | 50.1 | 69.7 | 79.6 | 3.025 | 9.543 | — | — | — |