fairseq2-feedstock

A conda recipe for fairseq2. fairseq2 is a sequence modeling toolkit that allows researchers and developers to train custom models for translation, summarization, language modeling, and other content generation tasks. It is also the successor of fairseq.

Home:

https://github.com/facebookresearch/fairseq2

Feedstock license:

BSD-3-Clause

Upstream license:

MIT

Docs:

https://facebookresearch.github.io/fairseq2/stable

Dev:

https://github.com/facebookresearch/fairseq2