Motion Captioning on HumanML3D
69.9BLEU@1Ours-[adapt](0,3)
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Ours-[adapt](0,3)Search strategy=Greedy Search, lambda_spat=0, lambda_adapt=32023.10 | 69.9 | 25 | 55.3 | 61.6 | 40.3 | |
| Ours-[adapt](0,3)Search strategy=Beam Search, Beam size=2, lambda_spat=0, lambda_adapt=32023.10 | 67.9 | 25.5 | 54.7 | 64.6 | 43.2 | |
| MLP+GRUSearch strategy=Beam Search, Beam size=22023.10 | 67 | 23.4 | 53.8 | 53.7 | 37.2 | |
| TM2TSearch strategy=Beam Search, Beam size=22023.10 | 61.7 | 22.3 | 49.2 | 72.5 | 37.8 | |
| SeqGANSearch strategy=Beam Search, Beam size=22023.10 | 47.8 | 13.5 | 39.2 | 50.2 | 23.4 |