microsoft/DeepSpeed

Pull Requests Commits

support large hidden dimensions (4K and 8K)

Reza Yazdani committed 4 years ago

9d2b71a5

supporting arbitrary hidden dimension for transformer kernels

Reza Yazdani committed 4 years ago

f850cc35

add adamW and deepspeed-adamw mode

Reza Yazdani committed 4 years ago

e2f204a3

supporting cpu-adam optimizer for zero-offload on deepspeed side

Reza Yazdani committed 4 years ago

283da002

add adamW to CPU-ADAM implementation

Reza Yazdani committed 4 years ago

6faa75b1

updating website dependencies (#475)

Shaden Smith committed 4 years ago

Verified d720fdb6

Add support for p100 in transformer kernels (#470)

jeffra committed 4 years ago

Verified 7ddfda85

revert previous (accidental) change

jeffra committed 4 years ago

Verified 1afca8f7

add compute cap of 6.0 to transformer kernels

jeffra committed 4 years ago

Verified b8eb40eb

fixing typo (#460)

RezaYazdaniAminabadi committed 4 years ago

Verified e25f2a23

Add DeepSpeed_Adam optimizer (#468)

tjruwase committed 4 years ago

Verified 23fc48f3

Fix printing momentum for non-deepspeed optimizer (#464)

Zixxy committed 4 years ago

Verified c39a76fe

gan tutorial (#462)

niumanar committed 4 years ago

Verified 2efea694

turning off different tests (temp)

jeffra committed 4 years ago

679fc135

temporarily disable lr unit tests

jeffra committed 4 years ago

11cf47ef

Update engine.py (#458)

Zixxy committed 4 years ago

Verified 67176383

Use parentesis around min and max to enable Windows build (#449)

bratao committed 4 years ago

Verified 95575579

Disable default installation of CPU Adam (#450)

tjruwase committed 4 years ago

Verified 7b8be2a7

fix typos (#446)

eric-haibin-lin committed 4 years ago

Verified 6f28ea30

unit test rename (#442)

Shaden Smith committed 4 years ago

Verified 5412a334

link fix part two :-) (#441)

Shaden Smith committed 4 years ago

Verified 6d176c45

Update pipeline.md (#439)

eric-haibin-lin committed 4 years ago

Verified 0ca82156

Update azure.md (#437)

conglongli committed 4 years ago

Verified 192cf7c8

Fix urls in tutorial (#436)

conglongli committed 4 years ago

Verified 5d40f006

support dynamic sequence length in transformer kernels (#424)

RezaYazdaniAminabadi committed 4 years ago

Verified f0f2a702

DSE bump (#427)

Shaden Smith committed 4 years ago

Verified 71f7df39

Add configurable intermediate size to transformer kernels (#423)

RezaYazdaniAminabadi committed 4 years ago

Verified a148bd33

Fix activation checkpoint unit tests for GPU systems (#421)

Shaden Smith committed 4 years ago

Verified a825f996

Revert "Activation checkpointing bugfix and unit tests (#420)" (#422)

jeffra committed 4 years ago

Verified a74a604a

Activation checkpointing bugfix and unit tests (#420)

Shaden Smith committed 4 years ago

Verified 01b6e27e

Older