DeepSpeed
Universal Checkpointing: read AutoEP ZeRO-3 optimizer files one at a time
#8844
Merged

Universal Checkpointing: read AutoEP ZeRO-3 optimizer files one at a time #8844

pengdurice
pengdurice Universal Checkpointing: read AutoEP ZeRO-3 optimizer files one at a …
81a4919b
pengdurice Merge master into peng-fix-ucp-stream-expert-states-v1
be499f3e
pengdurice Universal Checkpointing: free each rank's optimizer state before the …
d031286d
gpu-test-labeler gpu-test-labeler added skip-gpu-test
gpu-test-labeler gpu-test-labeler removed skip-gpu-test
gpu-test-labeler gpu-test-labeler added skip-gpu-test
pengdurice Merge branch 'master' into peng-fix-ucp-stream-expert-states-v1
3275550c
pengdurice pengdurice marked this pull request as ready for review 5 days ago
pengdurice pengdurice requested a review from tjruwase tjruwase 5 days ago
gpu-test-labeler gpu-test-labeler removed skip-gpu-test
chatgpt-codex-connector
gpu-test-labeler gpu-test-labeler added skip-gpu-test
chatgpt-codex-connector
chatgpt-codex-connector commented on 2026-11-14
tohtana
tohtana requested changes on 2026-11-16
pengdurice Universal Checkpointing: write AutoEP expert fragments below created …
38686244
pengdurice pengdurice requested a review from loadams loadams 3 days ago
gpu-test-labeler gpu-test-labeler removed skip-gpu-test
gpu-test-labeler gpu-test-labeler added skip-gpu-test
pengdurice
tohtana
tohtana approved these changes on 2026-11-17
tohtana tohtana merged 17828f6e into master 2 days ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone