Simultaneous Speech Translation on MuST-C EN-DE (tst-COMMON)
23.8BLEULS-Transducer-SST
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| LS-Transducer-SSTepsilon=5, AIF=true, Latency=Medium2024.06 | 23.8 | — | — | 1.635 | — | |
| TBCALatency=High2024.06 | 23.5 | — | — | 2.3 | — | |
| LS-Transducer-SSTepsilon=3, AIF=true, Latency=Medium2024.06 | 23.3 | — | — | 1.188 | — | |
| Wait-kk=5, step_size=360 ms, Latency=High2024.06 | 22.5 | — | — | 2.629 | — | |
| Wait-kk=3, step_size=360 ms, Latency=High2024.06 | 22.3 | — | — | 2.127 | — | |
| LS-Transducer-SSTepsilon=1, AIF=true, Latency=Low2024.06 | 21.4 | — | — | 0.853 | — | |
| Wait-kk=3, step_size=280 ms, Latency=Medium2024.06 | 21 | — | — | 1.638 | — | |
| LS-Transducer-SSTLatency=Low2024.06 | 20.8 | — | — | 0.715 | — | |
| Wait-kk=1, step_size=280 ms, Latency=Medium2024.06 | 20.6 | — | — | 1.28 | — | |
| CAATLatency=Medium2024.06 | 20.4 | — | — | 1.078 | — | |
| MoSSTLatency=High2024.06 | 20 | — | — | 2.742 | — | |
| MoSSToperating_point=52021.09 | 19.97 | 4,278 | 96 | 2,742 | — | |
| MoSSTk=inf2021.09 | 19.95 | — | — | — | — | |
| MoSSToperating_point=42021.09 | 19.77 | 3,827 | 93 | 2,551 | — | |
| MoSSTk=92021.09 | 19.37 | — | — | — | — | |
| CIF-ILlambda_lat=0, Latency=Medium2024.06 | 19.3 | — | — | 1.354 | — | |
| MoSSTk=72021.09 | 18.71 | — | — | — | — | |
| Wait-k2024.06 | 18.6 | — | — | 6.8 | — | |
| CIF-ILlambda_lat=0.3, Latency=Low2024.06 | 18.5 | — | — | 0.962 | — | |
| SimulSpeechk=inf2021.09 | 18.29 | — | — | — | — | |
| MoSSTk=52021.09 | 18.23 | — | — | — | — | |
| SimulSpeechk=92021.09 | 17.87 | — | — | — | — | |
| SimulSpeechk=72021.09 | 17.46 | — | — | — | — | |
| Wait-kk=3, step_size=200 ms, Latency=Medium2024.06 | 17 | — | — | 1.014 | — | |
| SimulSpeechk=52021.09 | 16.9 | — | — | — | — | |
| Wait-kk=2, step_size=200 ms, Latency=Low2024.06 | 16.5 | — | — | 0.881 | — | |
| MoSSToperating_point=32021.09 | 16.34 | 2,263 | 79 | 1,734 | — | |
| Wait-kk=1, step_size=200 ms, Latency=Low2024.06 | 16.3 | — | — | 0.757 | — | |
| SimulSToperating_point=52021.09 | 15.99 | 5,271 | 99 | 3,049 | — | |
| MoSSTk=32021.09 | 15.57 | — | — | — | — | |
| SimulSpeechk=32021.09 | 15.52 | — | — | — | — | |
| SimulSToperating_point=42021.09 | 14.17 | 4,361 | 96 | 2,774 | — | |
| MoSSTk=12021.09 | 11.76 | — | — | — | — | |
| SimulSToperating_point=32021.09 | 10.88 | 2,946 | 88 | 2,165 | — | |
| SimulSpeechk=12021.09 | 10.73 | — | — | — | — | |
| MoSSToperating_point=22021.09 | 6.75 | 1,182 | 53 | 818 | — | |
| SimulSToperating_point=22021.09 | 3.6 | 1,500 | 62 | 1,097 | — | |
| MoSSToperating_point=12021.09 | 1.35 | 642 | 29 | 208 | — | |
| SimulSToperating_point=12021.09 | 0.25 | 930 | 37 | 604 | — | |
| SimulUAveraged across=cutoff frame f2026.03 | — | — | — | 247 | 137 | |
| Top CascadeAveraged across=speech segment size2026.03 | — | — | — | 246 | 137 |