Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

activation ordering #2316

Open
wants to merge 15 commits into
base: main
Choose a base branch
from
Open

activation ordering #2316

wants to merge 15 commits into from

Conversation

horheynm
Copy link
Member

@horheynm horheynm commented Jun 4, 2024 •
edited
Loading

Activation Ordering implementation.

Checked lm_eval value with actorder=True

"metrics":[{"name":"word_perplexity,none","value":10.17568878732032},

@horheynm horheynm marked this pull request as ready for review June 5, 2024 18:22
@robertgshaw2-neuralmagic
Copy link
Contributor

robertgshaw2-neuralmagic commented Jun 5, 2024 •
edited
Loading

As it relates to the serialization of act_reordering in compressed-tensors, @dsikka can you make sure this works well with the code we have in vllm?

[ note: I know this is unrelated to this PR ]

dbogunowicz
dbogunowicz previously approved these changes Jun 6, 2024
View reviewed changes
Copy link
Contributor

@dbogunowicz dbogunowicz left a comment

Choose a reason for hiding this comment

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

Could we have some (manual) tests in place for this as well? It's a subtle modification of the behavior, it would be great to have tested one way or another.

src/sparseml/modifiers/utils/layer_compressor.py Outdated Show resolved Hide resolved
src/sparseml/modifiers/quantization/gptq/utils/gptq_wrapper.py Outdated Show resolved Hide resolved
src/sparseml/modifiers/quantization/gptq/pytorch.py Outdated Show resolved Hide resolved
Satrat
Satrat requested changes Jun 17, 2024
View reviewed changes
Copy link
Contributor

@Satrat Satrat left a comment

Choose a reason for hiding this comment

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

Could we add a unit test that tests the perplexity for oneshot done with this flag?

src/sparseml/modifiers/quantization/gptq/base.py Show resolved Hide resolved
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@Satrat Satrat Satrat requested changes

@dbogunowicz dbogunowicz dbogunowicz left review comments

Requested changes must be addressed to merge this pull request.

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
4 participants
@horheynm @robertgshaw2-neuralmagic @Satrat @dbogunowicz