论文信息 - Sparse Upcycling: Training Mixture-of-Experts from Dense Checkpoints - 字舞流文

Sparse Upcycling: Training Mixture-of-Experts from Dense Checkpoints

Carlos Riquelme Ruiz | N. Houlsby | M. Dehghani | J. Ainslie | Aran Komatsuzaki | Yi Tay | Basil Mustafa | J. Puigcerver | J. Lee-Thorp | Mostafa Dehghani