Models With Tied Weights Need Re-Tieing After FSDP Param Init #3154
add fsdp_tool to retie after param init
5c53a83f
fabianlim
changed the title Models With Tied Weights Need Re-Tieing After Models With Tied Weights Need Re-Tieing After FSDP Param Init 1 year ago
make it handle generic param_init_fn
50b71274
muellerzr
approved these changes
on 2024-10-31
fix quality
886ba0ef
muellerzr
merged
8159c98d
into main 1 year ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub