Temporal Action Detection on Charades
57.2mAP@0.2ESTF-SSM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ESTF-SSMBackbone=VideoMAEv2 [32], E2E=true2026.04 | 57.2 | 45.5 | 27.1 | 38.9 | |
| MambaTADBackbone=VideoMAEv2 [32], E2E=true2026.04 | 54.1 | 43.2 | 26.6 | 37.7 | |
| AdaTADBackbone=VideoMAEv2 [32], E2E=true2026.04 | 53.7 | 42.9 | 26.7 | 37.5 | |
| AANBackbone=CLIP [37], E2E=false2026.04 | 51.2 | 38.4 | 20.7 | 32 | |
| MT-TembaBackbone=CLIP [37], E2E=false2026.04 | 51.2 | 38.4 | 20.7 | 32.3 | |
| MS-TSTBackbone=CLIP [37], E2E=false2026.04 | 50.1 | 37.9 | 19.2 | 31.9 |