microsoft/DeepSpeed

Pull Requests Commits

Merge branch 'master' into lokoppak/ln_schedule_update

cmikeh2 committed 2 years ago

Verified 01080fc3

Migrate ops tests to new inference_ops marker (#2599)

cmikeh2 committed 2 years ago

Verified 7425a365

fix blog link (#2600)

conglongli committed 2 years ago

Verified acde873c

DeepSpeed Data Efficiency Library (#2585)

conglongli committed 2 years ago

Verified ef869377

Merge branch 'master' into lokoppak/ln_schedule_update

cmikeh2 committed 2 years ago

Verified ea44e604

jeffra committed 2 years ago

2600db54

lokoppakmsft committed 2 years ago

edd17bda

move pow2 function

lokoppakmsft committed 2 years ago

4dd969d2

revert unwanted file

lokoppakmsft committed 2 years ago

49b96e10

Merge branch 'master' into lokoppak/ln_schedule_update

cmikeh2 committed 2 years ago

Verified 85b28716

get mask token from tokenizer (#2592)

jeffra committed 2 years ago

Verified 2076bf23

lokoppakmsft committed 2 years ago

85f06f8b

lokoppakmsft committed 2 years ago

ca3a0f5a

lokoppakmsft committed 2 years ago

1e6a2cd8

lokoppakmsft committed 2 years ago

48fef51a

lokoppakmsft committed 2 years ago

bddbce6b

lokoppakmsft committed 2 years ago

53af4cec

Pre-commit fixes

lokoppakmsft committed 2 years ago

82be5cb4

Fix issues w. python 3.6 + add py-version checks to CI (#2589)

jeffra committed 2 years ago

Verified 35eabb0a

Move layer norm to new schedule

lokoppakmsft committed 2 years ago

6029242f

Updating API docs (#2586)

jomayeri committed 2 years ago

Verified 18713c68

Updating docs README (#2587)

jomayeri committed 2 years ago

Verified 377c770a

Add checkpoint sharding unit tests (#2561)

mrwyattii committed 2 years ago

Verified ccb8eb81

Support N-dimension input in quantization kernel (#2575)

lokoppakmsft committed 2 years ago

Verified 591744eb

Update barrier and reduce_scatter_base to conform to PyTorch signatures (#2570)

Quentin-Anthony committed 2 years ago

Verified 18d55e54

Fix MegatronLayerPolicy to have megatron_v2=True (#2579)

lekurile committed 2 years ago

Verified 731965db

Fix quantized-inference & Add generic support of checkpoint loading (#2547)

RezaYazdaniAminabadi committed 2 years ago

Verified 35b350b2

Drop Maxwell Support (#2574)

cmikeh2 committed 2 years ago

Verified b8416282

Support fp32 gradaccum for bf16 model (#2566)

delock committed 2 years ago

Verified 06938835

Add Determined to open-source DL frameworks (#2573)

sirredbeard committed 2 years ago

Verified 2d8f3f56

Older