Commits microsoft/DeepSpeed

Point the installer's error messages at the renamed branch

sfc-gh-truwase committed 2 days ago

f4fdcb6b

Run the packed tiled-remat installer tests against the DeepSpeed module

sfc-gh-truwase committed 2 days ago

09035751

Accept shards in install_packed_seq_tiled_rematerialize

sfc-gh-truwase committed 2 days ago

89c01f49

Add Hugging Face integration for packed-sequence tiled rematerialization

sfc-gh-truwase committed 2 days ago

cd61922c

Raise CPU checkpoint buffer pool count default; add opt-in live-bytes probe

sfc-gh-truwase committed 2 days ago

07d1ae54

Add packed-sequence tiled rematerialization with activation offload.

sfc-gh-truwase committed 2 days ago

0b092965

[Hybrid Engine] Require an explicit accelerator opt-in for graph capture (#8980)

VenusTZZ committed 3 days ago

Verified b62afa82

Check that converting through an affine map writes every element once (#8975)

Achyuthan-S committed 3 days ago

Verified 7e58feb6

Clear stale offload scratch before restoring low-precision gradients (#8728)

adenzhou1350 committed 3 days ago

Verified 0d0aa30b

Fix AutoEP storage-map transfer test for CPU tensor views (#9012)

jinyouzhi committed 3 days ago

Verified 67c83616

Support Transformers 5.20 cache query offsets in DeepSpeedStaticCache (#8852)

nathon-lee committed 3 days ago

Verified a23268d0

Fix bug in Adagrad optimizer initialization (#8976)

hwchen2017 committed 3 days ago

Verified 735c4754

Count each expert once in the fp32 gradient-clipping norm (#8478)

alanhuangyoo committed 3 days ago

Verified 31e6abca

Fix torch-latest workflow unit failures (#8950)

sfc-gh-truwase committed 3 days ago

Verified a38957c1

[AutoTP] Support frozen AutoTP models and manage AutoTP mode per model (#8851)

jinyouzhi committed 4 days ago

Verified d8241a9b

XPU: register pinned host memory via Level Zero ext_memmap (#8887)

delock committed 4 days ago

Verified 36b2ab3b

[checkpoint] Execute affine transfer plans with bounded peer staging (#8809)

0z5a committed 4 days ago

Verified f3c6cbb2

ZeRO-3: build the partial-offload device string from the public accelerator API (#8994)

GuCarpenter committed 4 days ago

Verified d9225000

Keep float32 intermediates in the clamped SwiGLU eager fallback (#8992)

GuCarpenter committed 4 days ago

Verified 961df84b

Add an opt-in deterministic execution policy (#8917)

ziming-zh committed 4 days ago

Verified dfc63adf

Update version.txt post-v0.19.9 release (#8999)

loadams committed 4 days ago

Verified 2cd4984a

ZeRO-1/2: preserve gradients across request boundaries until step (#8949)

stas00 committed 5 days ago

Verified dc046b15

[VocabTP] vocab_parallel_ce support fused LM-head projection + loss for the HF forward (#8945)

jinyouzhi committed 6 days ago

Verified 1a956270

Cap CPU accelerator memory at the cgroup limit of the process (#8930)

prabhkodes committed 6 days ago

Verified 2d80cedd

[Hybrid Engine] Enable graph capture for HybridEngine on non-CUDA accelerators (#8933)

VenusTZZ committed 6 days ago

Verified ab7fd98f

ci: pin torch<2.15 for MPS workflow until pytorch/pytorch#202529 is fixed (#8955)

delock committed 7 days ago

Verified ba0d63ce

ZeRO-3: copy offloaded master weights back to the device through pinned buffer (#8840)

pengdurice committed 8 days ago

Verified c6291b7e

Add klxpu (KUNLUNXIN XPU) self-hosted CI for the v1 unit suite (#8923)

GuCarpenter committed 9 days ago

Verified 27360990

[AffineIR] Storage maps of AutoEP experts under EDP and ZeRO (#8941)

jinyouzhi committed 9 days ago

Verified 66fef47b

Sum GPTBigCode's shared key/value gradient across tensor-parallel ranks (#8928)

Achyuthan-S committed 9 days ago

Verified 9148ab22