Machine Translation on PhoMT 1.0 (test)
45.47BLEUEnViT5 base
Evaluation results
| Method | Links | |
|---|---|---|
| EnViT5 base# Params=275M, Pretrained=CC100, Finetuned Dataset=MTet + PhoMT, Finetuned # pairs=6.2M2022.10 | 45.47 | |
| EnViT5 base# Params=275M, Pretrained=CC100, Finetuned Dataset=MTet, Finetuned # pairs=4.2M2022.10 | 43.87 | |
| mBART# Params=448M, Pretrained=CC25, Finetuned Dataset=PhoMT, Finetuned # pairs=3M2022.10 | 43.46 | |
| Transformer-big# Params=213M, Finetuned Dataset=PhoMT, Finetuned # pairs=3M2022.10 | 42.94 | |
| Transformer-base# Params=65M, Finetuned Dataset=PhoMT, Finetuned # pairs=3M2022.10 | 42.12 | |
| EnViT5 base# Params=275M, Pretrained=CC100, Finetuned Dataset=MTet + PhoMT, Finetuned # pairs=6.2M2022.10 | 40.57 | |
| Bing Translator2022.10 | 40.37 | |
| Google Translate2022.10 | 39.86 | |
| mBART# Params=448M, Pretrained=CC25, Finetuned Dataset=PhoMT, Finetuned # pairs=3M2022.10 | 39.78 | |
| EnViT5 base# Params=275M, Pretrained=CC100, Finetuned Dataset=MTet, Finetuned # pairs=4.2M2022.10 | 39.57 | |
| Transformer-big# Params=213M, Finetuned Dataset=PhoMT, Finetuned # pairs=3M2022.10 | 37.83 | |
| Transformer-base# Params=65M, Finetuned Dataset=PhoMT, Finetuned # pairs=3M2022.10 | 37.19 | |
| M2M100# Params=1.2B, Finetuned Dataset=CCMatrix + CCAligned, Finetuned # pairs=7.5B2022.10 | 35.83 | |
| Google Translate2022.10 | 35.76 | |
| Bing Translator2022.10 | 35.74 | |
| M2M100# Params=1.2B, Finetuned Dataset=CCMatrix + CCAligned, Finetuned # pairs=7.5B2022.10 | 31.15 |