undefined | Better HN

0 pointsamelius5mo ago0 comments

Nor that the training from scratch will even work.

0 comments

1 comments · 1 top-level

exactly, that is the current objective. To proove that generation for a specific domain is on-par with causal attention models

j / k navigate · click thread line to collapse