GitLab

push v0.12.3 version

See merge request dcutoolkit/deeplearing/deepspeed!4

Ds v0.9.2 rocm

See merge request dcutoolkit/deeplearing/deepspeed!3

Ds v0.9.2 rocm

See merge request dcutoolkit/deeplearing/deepspeed!2

Ds v0.8.2 rocm

See merge request aicomponent/deepspeed!3

modify error

See merge request aicomponent/deepspeed!2

Ds v0.8.2 rocm, support torch1.13 for the hipify change

See merge request aicomponent/deepspeed!1

Merge branch 'deepspeed-0.6.3-rocm' of http://10.0.100.3/dcutoolkit/deeplearing/deepspeed into deepspeed-0.6.3-rocm
version modify

Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

This test has been giving us trouble for a bit, seeing nondeterministic failures, skipping for now to not break out CI. Need to revisit soon though.

* zero.Init() clarification

clarify that if `model.half()` can't fit into gpu memory `zero.Init()` is a must.

this proposal is via @samyam's clarification shared elsewhere.

Thank you.

* style

* add clarity

* style
Co-authored-by: Olatunji Ruwase <olruwase@microsoft.com>