Simultaneous Machine Translation on MuST-C en-de v1.0 (test)
27.32BLEUNMT LLM Chunk SBS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| NMT LLM Chunk SBSwait-k value=7, beams=5, chunks=4, window size=202023.12 | 27.32 | 6.97 | |
| NMT TransformerDecoding Scheme=non-simultaneous2023.12 | 26.96 | 22.6 | |
| NMT LLM Single SBSwait-k value=3, beams=5, chunks=1, window size=62023.12 | 25.98 | 4.12 | |
| NMT LLM Single SBSwait-k value=3, beams=5, chunks=1, window size=102023.12 | 25.95 | 4.25 | |
| NMT LLMDecoding Scheme=wait-32023.12 | 25.83 | 3.65 | |
| NMT LLM Chunk SBSwait-k value=5, beams=5, chunks=3, window size=152023.12 | 25.8 | 5.61 | |
| Simul-LLMFine-tuning wait-k=7, Inference wait-k=7, Training samples=5M2023.12 | 24.93 | 6.7 | |
| NMT LLM Chunk SBSwait-k value=3, beams=5, chunks=2, window size=102023.12 | 23.61 | 4.63 | |
| Simul-LLMFine-tuning wait-k=7, Inference wait-k=7, Training samples=2M2023.12 | 23.09 | 6.71 | |
| Monotonic TransformerWait-k=5, Decoding Scheme=SimulMT2023.12 | 22.01 | 3.32 | |
| Simul-LLMFine-tuning wait-k=7, Inference wait-k=3, Training samples=5M2023.12 | 21.86 | 3.38 | |
| Simul-LLMFine-tuning wait-k=7, Inference wait-k=3, Training samples=2M2023.12 | 20.82 | 3.44 | |
| Simul-LLMFine-tuning wait-k=3, Inference wait-k=3, Training samples=2M2023.12 | 19.99 | 3.41 |