Skip to content

Conversation

@stas00
Copy link
Contributor

@stas00 stas00 commented Sep 16, 2021

attempting to fix ci

@stas00 stas00 closed this Nov 18, 2021
adammoody pushed a commit to adammoody/Megatron-DeepSpeed that referenced this pull request Jun 21, 2023
* xpu support (bigscience-workshop#55)

* port accel abs interfece

* WA for run3.6b

* move on

* fix current_dievice

* fix typo

* enable to run 345M GPT

* delete apex_patch

* add TODO xpu compatible tg for xpu WA

* use deepspeed launcher

* enable run3.6b bf16

* add zero2 config json

* readd enable_each_rank_log

* fix typos

* add ccl arg

* fix

* use short word

* use no-masked-softmax-fusion

* readd

* set train  iters to 10

* remove duplicate line

* change assert msg

* update format

* add whitespace

* update path

* update note

* update

* fix typos

* delete notes

* update format

* update xpu check to cuda check

* update

* clean up file

* fix typos

* add python based gradient clipping

* change condition for python based path
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant