microsoft/DeepSpeed

Pull Requests Commits

Include torch in deps, no build isolation on v100 tests either

loadams committed 1 year ago

22900d34

Include torch in deps, no build isolation

loadams committed 1 year ago

d9ebd0e7

Don't include torch in deps, no build isolation

loadams committed 1 year ago

3cdd5451

Add build to the pyproject

loadams committed 1 year ago

c4b24fd3

Remove unneeded requires in build system declaration

loadams committed 1 year ago

4c32a9d0

Merge branch 'master' into loadams/pyproject-toml

loadams committed 1 year ago

Verified e40df22d

Link AutoTP blog in the front page (#7167)

hwchen2017 committed 1 year ago

31ec2b77

[NFC] Typo fix in SP layer. (#7152)

c8ef committed 1 year ago

7b7ac9e3

Add destroy to tests to free memory (#7160)

tohtana committed 1 year ago

6adcf2e3

Avoid graph break by removing redundant requires_grad attr change (#7158)

deepcharm committed 1 year ago

4a851f35

hf tp+zero training doc. (#7151)

inkcherry committed 1 year ago

78ec0254

fix leak of z3 buffer

Masahiro Tanaka committed 1 year ago

2db922fe

Update container version that runs on A6000 tests. (#7153)

loadams committed 1 year ago

4ca7ba57

Enhance Gaudi2 CI/Nightly Coverage with Model Parallelism and Linear Tests (#7146)

raza-sikander committed 1 year ago

9d753116

Correct the BACKWARD_PREFETCH_SUBMIT mismatch (#7120)

A-transformer committed 1 year ago

ea61c6ca

Conditionally quote env vars (#7071)

saurabhkoshatwar committed 1 year ago

e0188ac3

Unpin transformers version for most workflows (#7139)

loadams committed 1 year ago

6de20f65

Add conditional expression (#7119)

A-transformer committed 1 year ago

8df6245e

Avoid missing attr error (#7133)

tjruwase committed 1 year ago

eaeae790

Update CONTRIBUTING.md to reflect changes from CLA to DCO (#7135)

loadams committed 1 year ago

e93f0d16

Training multiple models (#7018)

tjruwase committed 1 year ago

065084be

Add sequential pytest mark to TestNVMeCheckpointing to resolve pytest forked hangs (#7131)

loadams committed 1 year ago

62abd88e

fix keep_module_on_host (#7112)

inkcherry committed 1 year ago

c7934014

Update gaudi2 nightly,ci to latest 1.20.0 build (#7093)

raza-sikander committed 1 year ago

ec2e93e3

Update references to new X/Twitter handle (#7110)

loadams committed 1 year ago

1c35241b

Fix fused_qkv print model ValueError (#7109)

Yejing-Lai committed 1 year ago

240c2a7b

Avoid graph break due to unsupported frozenset (#7105)

deepcharm committed 1 year ago

674f9bd5

Only run pre-commit on the changes (#7106)

hwchen2017 committed 1 year ago

7b7e1122

Avoid graph breaks in torch.compile caused by inner classes in the backward hooks (#7062)

deepcharm committed 1 year ago

feeaa2e4

Avoid graph breaks by disabling sourceless calls in instrument_w_nvtx (#7081)

deepcharm committed 1 year ago

6eb5baa7

Older