Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2406.14553 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929582847819776 |
|---|---|
| author | Larionov, Daniil Seleznyov, Mikhail Viskov, Vasiliy Panchenko, Alexander Eger, Steffen |
| author_facet | Larionov, Daniil Seleznyov, Mikhail Viskov, Vasiliy Panchenko, Alexander Eger, Steffen |
| contents | State-of-the-art trainable machine translation evaluation metrics like xCOMET achieve high correlation with human judgment but rely on large encoders (up to 10.7B parameters), making them computationally expensive and inaccessible to researchers with limited resources. To address this issue, we investigate whether the knowledge stored in these large encoders can be compressed while maintaining quality. We employ distillation, quantization, and pruning techniques to create efficient xCOMET alternatives and introduce a novel data collection pipeline for efficient black-box distillation. Our experiments show that, using quantization, xCOMET can be compressed up to three times with no quality degradation. Additionally, through distillation, we create an 278M-sized xCOMET-lite metric, which has only 2.6% of xCOMET-XXL parameters, but retains 92.1% of its quality. Besides, it surpasses strong small-scale metrics like COMET-22 and BLEURT-20 on the WMT22 metrics challenge dataset by 6.4%, despite using 50% fewer parameters. All code, dataset, and models are available online at https://github.com/NL2G/xCOMET-lite. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_14553 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | xCOMET-lite: Bridging the Gap Between Efficiency and Quality in Learned MT Evaluation Metrics Larionov, Daniil Seleznyov, Mikhail Viskov, Vasiliy Panchenko, Alexander Eger, Steffen Computation and Language State-of-the-art trainable machine translation evaluation metrics like xCOMET achieve high correlation with human judgment but rely on large encoders (up to 10.7B parameters), making them computationally expensive and inaccessible to researchers with limited resources. To address this issue, we investigate whether the knowledge stored in these large encoders can be compressed while maintaining quality. We employ distillation, quantization, and pruning techniques to create efficient xCOMET alternatives and introduce a novel data collection pipeline for efficient black-box distillation. Our experiments show that, using quantization, xCOMET can be compressed up to three times with no quality degradation. Additionally, through distillation, we create an 278M-sized xCOMET-lite metric, which has only 2.6% of xCOMET-XXL parameters, but retains 92.1% of its quality. Besides, it surpasses strong small-scale metrics like COMET-22 and BLEURT-20 on the WMT22 metrics challenge dataset by 6.4%, despite using 50% fewer parameters. All code, dataset, and models are available online at https://github.com/NL2G/xCOMET-lite. |
| title | xCOMET-lite: Bridging the Gap Between Efficiency and Quality in Learned MT Evaluation Metrics |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2406.14553 |