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.

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

[nnpackage] Update circle schema #13930

Merged
merged 1 commit into from
Sep 8, 2024

Conversation

hseok-oh
Copy link
Contributor

@hseok-oh hseok-oh commented Sep 4, 2024

This commit updates circle schema to 0.9

  • Introduce GGML_Q{X}_{Y} types for ggml block quantization
  • Introduce weight compression type and option

Related issue: #13743 #13521

@hseok-oh hseok-oh added the PR/ready for review It is ready to review. Please review it. label Sep 4, 2024
@hseok-oh
Copy link
Contributor Author

hseok-oh commented Sep 4, 2024

New types are renamed to GGML_... based on #13743 (comment)

glistening
glistening previously approved these changes Sep 4, 2024
Copy link
Contributor

@glistening glistening left a comment

Choose a reason for hiding this comment

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

LGTM

@hseok-oh hseok-oh changed the title [nnpackage] Introduce ggml quantization types [nnpackage] Update circle schema Sep 5, 2024
@hseok-oh
Copy link
Contributor Author

hseok-oh commented Sep 5, 2024

@SlavikMIPT Please check weight compression

SlavikMIPT
SlavikMIPT previously approved these changes Sep 5, 2024
Copy link
Contributor

@SlavikMIPT SlavikMIPT left a comment

Choose a reason for hiding this comment

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

LGTM, thank you

@hseok-oh
Copy link
Contributor Author

hseok-oh commented Sep 6, 2024

@seanshpark @glistening PTAL

@@ -221,6 +230,12 @@ table VariantSubType {
has_rank: bool = false;
}

enum CompressionType : byte {
NONE = 0,
//Huffman encoding only
Copy link
Contributor

Choose a reason for hiding this comment

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

Suggested change
//Huffman encoding only
// Huffman encoding only

Copy link
Contributor

Choose a reason for hiding this comment

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

this comment seems unnecessary as it explains there is only Huffman ...

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I'll remove this comment

This commit updates circle schema to 0.9
- Introduce GGML_Q{X}_{Y} types for ggml block quantization
- Introduce weight compression type and option

ONE-DCO-1.0-Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
Copy link
Contributor

@seanshpark seanshpark left a comment

Choose a reason for hiding this comment

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

LGTM

Copy link
Contributor

@glistening glistening left a comment

Choose a reason for hiding this comment

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

LGTM

@seanshpark seanshpark merged commit 9213b12 into Samsung:master Sep 8, 2024
14 checks passed
@hseok-oh hseok-oh deleted the qx_y_type_circle branch September 9, 2024 01:44
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
PR/ready for review It is ready to review. Please review it.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants