llama: name the pre-rename PQ2_0 ftype instead of reporting unknown - #134
Merged
Conversation
ggufs packed before the Q2_0_G128 -> PQ2_0 rename carry the old ftype value 142. The format did not change, only the enum did, so these files load and compute correctly but print print_info: file type = unknown, may not work which reads as a broken file. With this they print print_info: file type = PQ2_0 - 2.13 bpw (group 128, legacy ftype) Naming only. No load path, no quantizer path, and no behaviour changes; the new constant exists so the switch has a case for the value. Worth noting the collision that makes this confusing: 142 is also GGML_TYPE_PQ2_0, the tensor type id, which is current and correct. Only the ftype was renumbered, to 141. Verified on a published gguf that carries ftype 142: reports unknown on the current branch, reports PQ2_0 with this change, same output either way.
khosravipasha
approved these changes
Aug 28, 2026
khosravipasha
left a comment
khosravipasha
left a comment
Collaborator
There was a problem hiding this comment.
LGTM
Should remove the extra one later the legacy one we can repack.
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Cosmetic, deferred from #121. Naming only.
What
ggufs packed before the
Q2_0_G128->PQ2_0rename carry the old ftype value 142. The format did not change, only the enum did, so those files load and compute correctly but report:which reads as a broken file. With this they report:
No load path, no quantizer path, and no behaviour change. The new constant exists so the name switch has a case for the value.
The confusing part
142 is also
GGML_TYPE_PQ2_0, the tensor type id, which is current and correct. Two different enums with adjacent values. Only the ftype was renumbered, to 141, and only the ftype is affected here.Verified
On a published gguf that carries ftype 142, same file both ways:
file type =unknown, may not workPQ2_0 - 2.13 bpw (group 128, legacy ftype)Output is identical either way, which is the point: these files were always fine, they just did not say so.