fairseq: A Fast, Extensible Toolkit for Sequence Modeling
About
fairseq is an open-source sequence modeling toolkit that allows researchers and developers to train custom models for translation, summarization, language modeling, and other text generation tasks. The toolkit is based on PyTorch and supports distributed training across multiple GPUs and machines. We also support fast mixed-precision training and inference on modern GPUs. A demo video can be found at https://www.youtube.com/watch?v=OtgDdWtHvto
Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli• 2019
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Machine Translation | WMT English-German 2014 (test) | BLEU29.3 | 136 | |
| Machine Translation | IWSLT German-to-English '14 (test) | BLEU Score33.95 | 110 | |
| Machine Translation | Hindi Visual Genome WAT 2021 (challenge set) | BLEU Score51.66 | 8 | |
| Machine Translation | Hindi Visual Genome WAT 2021 (test) | BLEU0.4412 | 8 | |
| Long-form Question Answering | ELI5 standard original | RL Score26.8 | 5 | |
| Image Classification | ImageNet 2009 | Top-1 Accuracy70.36 | 3 |
Showing 6 of 6 rows