Show HN: Recurser lib reduces GPT2-XL VRAM usage by 25% and runs it on Colab (opens in new tab)

(github.com)

5 pointshomo_sapiens3y ago1 comments

1 comments

1 comments · 1 top-level

Hi I'm the maker and I often imagine that in the future there will be a lot of medium-sized AIs that we can use in addition to super-sized AIs that we can call on to run continuously to handle tasks for us.

j / k navigate · click thread line to collapse