The new version update lacks training documentation for machine learning.

[hui-li-xf]

After the update, documentation on training machine translation models is missing. Does the fairseq2 framework support training NLLB-style multilingual translation models?