Making large AI models cheaper, faster and more accessible
[zero]support zero2 with gradient accumulation (#4511)
* support gradient accumulation with zero2 * fix type
L
LuGY committed
839847b7d78bce6af5dfe58d27b5ce2c74a3619b
Parent: c0efc3e
Committed by GitHub <noreply@github.com>
on 8/25/2023, 5:44:07 AM