microsoft/DeepSpeed

Pull Requests Commits

Merge branch 'master' into HeyangQin/fastgen_moe_h100

loadams committed 253 days ago

Verified 1bbfaacc

Pin nv-a6000 workflow (#6938)

loadams committed 253 days ago

Verified 66d3d3e9

Remove op compilation flags due to perf issue (#6944)

NirSonnenschein committed 253 days ago

Verified 396f8db7

Support pure meta model lm_head tp (#6812)

Yejing-Lai committed 256 days ago

Verified fa8db5cf

Add information on security expectations with this software (#6941)

loadams committed 257 days ago

Verified 1d15ef0a

Add position_ids arg to OPTEmbedding forward function (#6939)

lekurile committed 257 days ago

Verified 0fc3daad

Add deepseek autotp (#6937)

Yejing-Lai committed 257 days ago

Verified 45fce45c

Fix windows blog examples (#6934)

loadams committed 258 days ago

Verified 53fb5795

Fix building on Windows with presence of Triton (#6749)

woct0rdho committed 258 days ago

Verified b62c84d8

Update python version classifiers (#6933)

loadams committed 258 days ago

Verified 6628127a

Use `torch.log1p` (#6930)

kit1980 committed 259 days ago

Verified c41b0c28

inference: remove unused _validate_args function (#5505)

nelyahu committed 259 days ago

Verified c7f30322

Check transformers version in BLOOM for inference v1 (#6766)

lekurile committed 260 days ago

Verified f2cc8090

Cleanup ops/transformer/inference tests (#6925)

loadams committed 260 days ago

Verified c348c5b1

Reduce the device bubble introduced by heavy loop synchronization in coalesced fetch/release(z3_leaf_module) (#6694)

inkcherry committed 260 days ago

Verified b0040b6c

Add fp8_gemm fallback for non-triton systems (#6916)

oelayan7 committed 260 days ago

Verified c5e48f49

[BUG FIX]:fix get torch.version.cuda error when cuda is None in rocm (#6909)

kairos-yu committed 260 days ago

Verified f8c9f314

Fix `checkpointable_layers` Logic (#6881)

Quentin-Anthony committed 263 days ago

Verified 0dbbb70b

Cleanup ops/transformer/inference tests (#6830)

loadams committed 263 days ago

Verified a8ede3a9

Stage3: Use new torch grad accumulation hooks API (#6773)

deepcharm committed 264 days ago

Verified 456c9ac6

Change compile for pipeline module torch.compile (#6478)

NirSonnenschein committed 267 days ago

Verified 3573858e

Update Gaudi2 jobs to latest 1.19 build (#6905)

raza-sikander committed 271 days ago

Verified cc03c76d

Fix error caused by all_reduce call in domino (#6880)

hwchen2017 committed 271 days ago

Verified 85cc5f9b

hpu_accelerator: use torch.use_deterministic_algorithms (#6897)

nelyahu committed 278 days ago

Verified eea53048

Zero2: avoid graph breaks in torch.compile by using param_idx (#6803)

nelyahu committed 278 days ago

Verified 00ea0c46

Allow to compile collective for PT>2.3 (#6899)

NirSonnenschein committed 278 days ago

Verified 4fd79205

Update version.txt after 0.16.2 release (#6893)

loadams committed 279 days ago

Verified f9e158a0

Update code owners (#6890)

tjruwase committed 279 days ago

Verified b344c04d

Add arctic model support by adding w2 to all_reduce (#6856)

pi314ever committed 279 days ago

Verified 0b25630a

Don't error out when cpu accelerator doesn't have torch (as default for whl building) (#6886)

loadams committed 280 days ago

Verified 4cd1d974

Older