Refactor core_model_loading to support FSDP shard-on-read loading #44974
Refactor core_model_loading to support FSDP shard-on-read loading
4b2a9216
3outeille
force pushed
from
e5fc7eb3
to
4b2a9216
8 days ago
Merge fsdp-vs-ddp into fsdp-core-model-loading
c9f30d9f
Merge fsdp-vs-ddp (ruff fixes) into fsdp-core-model-loading
b7c3dc5c
linting
dfdfbdd9
from_pretrained distributed refactor (FSDP2 + TP) (#44996)
187ee5d4
DTensor-based TP + FSDP2 shard-on-read composability
607cc114
3outeille
force pushed
from
d48fcc7e
to
607cc114
6 days ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub