[1904.01038] fairseq: A Fast, Extensible Toolkit for Sequence Modeling

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

1 mentions: @r9y9
Keywords: fairseq
Date: 2019/07/06 02:17

Referring Tweets

@r9y9 t.co/pXjrC83jDA fairseqのpaper出てるの最近まで知らなかったけど、extensibleに設計されてるのが良いよなあ。DeepVoice3の実装したときにコード参考にしたのがもう二年前で、それから大分進化してる。Pytorchベースだしapexでmix precision 学習/推論もサポートしてる。良さみがあるなあ

Related Entries

Read more [1906.09573] A Neural Vocoder with Hierarchical Generation of Amplitude and Phase Spectra for Statis...
0 users, 1 mentions 2020/05/05 03:50
Read more Parallel Wavegan: A Fast Waveform Generation Model Based on Generative Adversarial Networks with Mul...
0 users, 1 mentions 2020/05/05 09:51
Read more Speech Synthesis and Voice Conversion I | IEEE ICASSP 2020
0 users, 1 mentions 2020/05/06 08:21
Read more TRANSFERRING NEURAL SPEECH WAVEFORM SYNTHESIZERS TO MUSICAL INSTRUMENT SOUNDS GENERATION | IEEE ICAS...
0 users, 1 mentions 2020/05/11 11:21
Read more GitHub - hitachi-speech/EEND: End-to-End Neural Diarization
0 users, 1 mentions 2020/06/30 11:21