microsoft/DeepSpeed

Pull Requests Commits

jeffra committed 4 years ago

0890aafe

Fix partial recovery of sparse_tensor_module_names and dynamically check if gradient data is sparse (#1562)

Dipet committed 4 years ago

Verified 4bf4ab7a

Add autotuning news post (#1565)

cli99 committed 4 years ago

Verified bda3d0e6

[build] support cuda-11.5 (#1558)

stas00 committed 4 years ago

Verified fa8d6c0a

Update offload parameter names (#1536)

tjruwase committed 4 years ago

Verified 7567c76c

Autotuning (#1554)

cli99 committed 4 years ago

Verified 9caa74e5

Add documentation for bfloat16 (git commit 648f7bfa5009484b822064d0c28d377da6dd71a0) (#1516)

manuelciosici committed 4 years ago

Verified b7cc7c8e

Fix zinf none swapper (#1550)

tjruwase committed 4 years ago

Verified 488105eb

Add warmup_type arguments in WarmupLR and WarmupDecayLR (#1530)

skpig committed 4 years ago

Verified 76847f42

Fix sparse attention for small block-sizes (#1545)

RezaYazdaniAminabadi committed 4 years ago

Verified 3ed77304

Tensor-Parallelism general support (#1512)

RezaYazdaniAminabadi committed 4 years ago

Verified 9ce00a21

backward compatibility (#1549)

conglongli committed 4 years ago

Verified b16dd943

jeffra committed 4 years ago

fa9d3e84

Fix 1bit extra issue (#1542)

jeffra committed 4 years ago

Verified 2665c8b1

Use cuda tensors for allgather (#1548)

tjruwase committed 4 years ago

Verified bd3ebddf

CPU-Adam: Fix compile Issue (#1537)

RezaYazdaniAminabadi committed 4 years ago

Verified af443f63

Modify inference engine (#1520)

chunyang-wen committed 4 years ago

Verified f0122007

[unit tests] allow unique port for tests

jeffra committed 4 years ago

0af15b98

fstr for multnode_runner (#1532)

chunyang-wen committed 4 years ago

Verified 93c71831

[docs] fix 404 (#1531)

stas00 committed 4 years ago

Verified 76f2b5e5

typo in profiler.py (#1527)

ncfrey committed 4 years ago

Verified 2c62d657

Adding Tutel to MoE layer (#1528)

alexandremuzio committed 4 years ago

Verified 2887349c

Use fstr in launcher (#1521)

chunyang-wen committed 4 years ago

Verified cf1f1601

make conv flops counting general for 1,2,3d (#1518)

cli99 committed 4 years ago

Verified f9b37801

jeffra committed 4 years ago

426dd2b5

Prevent creation of local temp directory (#1494)

aphedges committed 4 years ago

Verified 91defd7c

Unify use f str (#1511)

chunyang-wen committed 4 years ago

Verified df5b0884

[code readability] pipe (#1510)

stas00 committed 4 years ago

Verified bf1725bb

Remove redundant pass (#1509)

chunyang-wen committed 4 years ago

Verified 85ce85dd

Bfloat16 zero2 (#1398)

raamjad committed 4 years ago

Verified 648f7bfa

Older