quant : do not require imatrix when the tensor keeps its type - #26255
Open
TrevorS wants to merge 1 commit into
Open
quant : do not require imatrix when the tensor keeps its type#26255TrevorS wants to merge 1 commit into
TrevorS wants to merge 1 commit into
Conversation
A tensor whose target type equals its current type is copied verbatim by the quantize loop (cur_type != new_type short-circuit), so the imatrix data is unused by construction. Only set requires_imatrix when the type actually changes, so that --tensor-type pins holding IQ-class tensors at their source type no longer abort a partial requantization with "this quantization requires an importance matrix!" (and no longer trigger the spurious --dry-run imatrix warning).
|
Hi @TrevorS, thanks for your contribution! Per our contribution guidelines, the automated PR checker found the following issue(s) that need your attention:
Please note that maintainers reserve the right to make final decisions on PRs. If you believe there is a mistake, please comment below. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Overview
While claude was requantizing the deepseek iq3xxs model while pinning the experts in place, it came across a regression. Minimal repro on any model that already contains an imatrix-requiring type, pinning that type in place:
aborts in the pre-quantization validation with
ERROR: this quantization requires an importance matrix!even though the pinned tensors are byte-copied rather than quantized (regression introduced with the early validation pass in #19770).To fix the regression you can avoid setting
requires_imatrixwhentarget_typeandtensor->typeare the same.Requirements
--dry-run, and a still-aborts control on a genuinely converting pin).