microsoft/DeepSpeed

Pull Requests Commits

Merge branch 'arpan/elasticity' into arpan/auto-check

Arpan Jain committed 2 years ago

fc3e62b6

auto-check and two-level prototype

Arpan Jain committed 2 years ago

adaa283b

removed extra imports

Arpan Jain committed 2 years ago

66205a12

fixed test_elastic

Arpan Jain committed 2 years ago

f2405bd3

Arpan Jain committed 2 years ago

6ed80660

resolved conflicts

Arpan Jain committed 2 years ago

bb4a7f3d

modified elastic training unit test, added config options for elastic training in docs, and added an assertion in runner for elastic training

Arpan Jain committed 2 years ago

f4ace715

fix: fix BF16_Optimizer compatibility issue with optimizer state 0-dim tensor (#2152)

shjwudp committed 2 years ago

Verified 57140e8e

Graceful exit on failures for multi-node runs (#2008)

Jerry Mannil committed 2 years ago

Verified 66d29b0a

Trajepl/nebula ckpt engine (#2085)

trajepl committed 2 years ago

Verified e669aaf5

force newer datasets version (#2147)

jeffra committed 2 years ago

Verified a54661a0

formatting fix for #1962

jeffra committed 2 years ago

b442264d

Add purely-local sliding window sparse attention config (#1962)

Quentin-Anthony committed 2 years ago

Verified be46ff6d

Refactor ZeRO configs to use Pydantic (#2004)

mrwyattii committed 2 years ago

Verified 59975896

fix broken links (#2144)

jimwu6 committed 2 years ago

Verified 6f7137cb

[docs] README formatting fix

jeffra committed 2 years ago

76a01d5d

[docs] update build pipeline badges

jeffra committed 2 years ago

Verified 56a0d186

[docs] add details about install reqs (#2143)

jeffra committed 2 years ago

Verified 84dbe7c6

Allow turning off loss scaling wrt GAS + update tput calculator (#2140)

jeffra committed 2 years ago

Verified 5bd09a8f

turn off time check for now (#2142)

jeffra committed 2 years ago

Verified ddd2113d

Update 2022-07-26-deepspeed-azure.md (#2138)

awan-10 committed 2 years ago

Verified d9ec8ef4

adding HF Accelerate+DS tests workflow (#2134)

pacman100 committed 2 years ago

Verified 9f3a540d

Fix conflict between Tutel and top-2 gate in MoE layer (#2053)

yetiansh committed 2 years ago

Verified 31582d77

Update 2022-07-26-deepspeed-azure.md (#2136)

awan-10 committed 2 years ago

Verified 0e49b197

Azure blog news item (#2135)

awan-10 committed 2 years ago

Verified 2ff1cad0

DS on Azure blog (#2133)

awan-10 committed 2 years ago

Verified cf587b82

Add flake8 to pre-commit checks (#2051)

aphedges committed 2 years ago

Verified 316c4a43

Delete Gemfile.lock (#2130)

jeffra committed 2 years ago

Verified 9c6cdecc

DeepSpeed Communication Profiling and Logging (#2012)

Quentin-Anthony committed 2 years ago

Verified 5349347b

[ROCm] Wrong command broke ROCm build. (#2118)

jpvillam-amd committed 2 years ago

Verified b6305d0e

Older