Closed marcromeyn closed 1 year ago
This PR introduces mixture-of-experts + PLE/CGC. With this we should be able to write a pytorch version of the multi-task blogpost.
https://nvidia-merlin.github.io/models/review/pr-1173
Goals :soccer:
This PR introduces mixture-of-experts + PLE/CGC. With this we should be able to write a pytorch version of the multi-task blogpost.