-
Notifications
You must be signed in to change notification settings - Fork 608
Force callers to manually provide GPTQ args #2795
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
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/2795
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit b7a6f62 with merge base d612c23 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This pull request was exported from Phabricator. Differential Revision: D55605859 |
Summary: Default values for quant are convenient, but have caused confusion. Since there are no active users of GPTQ, we want to pre-empt potential ambiguity by making the following fields required and manually specified: * Group Size * Calibration Limit * Calibration Sequence Length --- Note: Group Size's default value is untouched since it is utilized by int4 quant and has active callers Differential Revision: D55605859
2ce4b40
to
8a5f093
Compare
This pull request was exported from Phabricator. Differential Revision: D55605859 |
Summary: Default values for quant are convenient, but have caused confusion. Since there are no active users of GPTQ, we want to pre-empt potential ambiguity by making the following fields required and manually specified: * Group Size * Calibration Limit * Calibration Sequence Length --- Note: Group Size's default value is untouched since it is utilized by int4 quant and has active callers Differential Revision: D55605859
Summary: Default values for quant are convenient, but have caused confusion. Since there are no active users of GPTQ, we want to pre-empt potential ambiguity by making the following fields required and manually specified: * Group Size * Calibration Limit * Calibration Sequence Length --- Note: Group Size's default value is untouched since it is utilized by int4 quant and has active callers Differential Revision: D55605859
8a5f093
to
d13164c
Compare
Summary: Default values for quant are convenient, but have caused confusion. Since there are no active users of GPTQ, we want to pre-empt potential ambiguity by making the following fields required and manually specified: * Group Size * Calibration Limit * Calibration Sequence Length --- Note: Group Size's default value is untouched since it is utilized by int4 quant and has active callers Reviewed By: mergennachin Differential Revision: D55605859
This pull request was exported from Phabricator. Differential Revision: D55605859 |
Summary: Default values for quant are convenient, but have caused confusion. Since there are no active users of GPTQ, we want to pre-empt potential ambiguity by making the following fields required and manually specified: * Group Size * Calibration Limit * Calibration Sequence Length --- Note: Group Size's default value is untouched since it is utilized by int4 quant and has active callers Reviewed By: mergennachin Differential Revision: D55605859
d13164c
to
b7a6f62
Compare
This pull request was exported from Phabricator. Differential Revision: D55605859 |
This pull request has been merged in 31d5c61. |
Summary:
Default values for quant are convenient, but have caused confusion.
Since there are no active users of GPTQ, we want to pre-empt potential ambiguity by making the following fields required and manually specified:
Note: Group Size's default value is untouched since it is utilized by int4 quant and has active callers
Differential Revision: D55605859