Skip to content

Add deepspeed for the glue finetuning tasks & bert-large scripts#77

Merged
conglongli merged 2 commits into
masterfrom
reyazda/glue-bertlarge-deepspeed
Jan 15, 2021
Merged

Add deepspeed for the glue finetuning tasks & bert-large scripts#77
conglongli merged 2 commits into
masterfrom
reyazda/glue-bertlarge-deepspeed

Conversation

@RezaYazdaniAminabadi
Copy link
Copy Markdown
Contributor

…and configs

Copy link
Copy Markdown
Contributor

@minjiazhang minjiazhang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me overall. Added one minor comment that needs to be addressed.

action='store_true',
default=False,
help=
"Whether to display the breakdown of the wall-clock time for foraward, backward and step"
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The description appears to be wrong. Perhaps change it "Switching to the variant of Transformer blocks that use pre-LayerNorm."

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Of course, thanks Minjia :)

Copy link
Copy Markdown
Contributor

@minjiazhang minjiazhang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM! Thanks for making the changes.

@conglongli conglongli merged commit 400cd1b into master Jan 15, 2021
hwchen2017 pushed a commit that referenced this pull request Jun 8, 2025
* Add deepspeed for the glue finetuning tasks & add bert-large scripts and configs

* change preln argument description
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.

3 participants