microsoft/DeepSpeed

Pull Requests Commits

fix the inference injection policies & add some description on the inference-engine parameters

Reza Yazdani committed 3 years ago

aceaa5b3

[docs] switch to transparent dark logo

jeffra committed 3 years ago

3a4cb042

fix the half-precision version of rotary_pos_emb kernel (#1683)

RezaYazdaniAminabadi committed 3 years ago

Verified 762e697a

GPT-J inference support (#1670)

RezaYazdaniAminabadi committed 3 years ago

Verified 289c3f9b

[docs] add gh-dark-mode logo

jeffra committed 3 years ago

7e857aab

[docs] add light-mode logo

jeffra committed 3 years ago

Verified 9c5cf3a5

add segfault guard for cpu-adam/adagrad (#1681)

jeffra committed 3 years ago

Verified 2422ec48

Copy grads to cpu in z1-offload (#1679)

tjruwase committed 3 years ago

Verified cef116f8

[docs] add logo (#1676)

jeffra committed 3 years ago

Verified c2735996

update results about public Pile dataset (#1675)

conglongli committed 3 years ago

Verified aca64799

Fix largest param numel calculation (#1623)

tjruwase committed 3 years ago

Verified 4354c3cc

support CPU Adam and Adagrad on Windows with SDK 10.0.22000 (#1634)

lostmsu committed 3 years ago

Verified 74493b2b

jeffra committed 3 years ago

b6f0ac97

Various small documentation text improvements (#1665)

manuelciosici committed 3 years ago

Verified d0ab7224

Convert the fp16_params to group of parameters (#1651)

RezaYazdaniAminabadi committed 3 years ago

Verified 559c4ce1

Replace calls to print() with calls to logger (#1664)

manuelciosici committed 3 years ago

Verified 40ce131c

[save_fp16_model] return status (#1663)

stas00 committed 3 years ago

Verified 317400ea

follow-up to #1652, resolved a100-80gb issue (#1655)

jeffra committed 3 years ago

Verified d93d924a

add backup cpu-arch detection if py-cpuinfo fails (#1652)

jeffra committed 3 years ago

Verified cbd68dc4

New feature contribution guideline (#1646)

conglongli committed 3 years ago

Verified 752319c7

Add torchvision to requirements-dev.txt (#1642)

aphedges committed 3 years ago

Verified 8bbf081a

Updating autotuner readme file to add hyperparameter adjustment suggestions (#1641)

minjiaz committed 3 years ago

Verified f2c433f0

Fix cpu-adam AVX performance (#1637)

RezaYazdaniAminabadi committed 3 years ago

Verified 259936a7

Add tensor methods in flops counting and separate macs and flops (#1591)

cli99 committed 3 years ago

Verified 082f392a

[testing] 3x faster unit tests (#1636)

jeffra committed 3 years ago

Verified 7f58853c

Refactor ZeRO naming to reduce confusion (#1607)

jeffra committed 3 years ago

Verified 1d295ff5

sharded_moe: make top1gating ONNX-exportable (#1578)

garymm committed 3 years ago

Verified 07887f66

use py-cpuinfo to detect SIMD_WIDTH in platform-independent way (#1616)

lostmsu committed 3 years ago

Verified 64c2946a

remove print (#1626)

awan-10 committed 3 years ago

Verified feb6afb0

MoE for NLG tutorial (#1633)

conglongli committed 3 years ago

Verified c6ace162

Older