Multilingual Machine Translation 2016 (dev)
3.35PerplexityMoE-Multi
Evaluation Results
| Method | Links | |
|---|---|---|
| MoE-MultiParameters=8.7B, ops/timestep=102M, training time=12 days, hardware=64 k40s2017.01 | 3.35 | |
| GNMT-MultiParameters=278M, ops/timestep=212M, training time=21 days, hardware=96 k20s2017.01 | 4.14 |