Repository navigation
model: support vllm decision 2 model - #30378
Merged
Merged
Conversation
github-actions
Bot
added
documentation
ggerganov
reviewed
Oct 12, 2026
| cur = build_head(cur, inp_decision); | ||
|
|
||
| // row i of the output is the score of option i | ||
| cur = ggml_pad(ctx0, cur, 0, n_outputs - cur->ne[1], 0, 0); |
Member
There was a problem hiding this comment.
Is this pad needed? Wouldn't we always have n_outputs == cur->ne[1] here?
Collaborator
Author
There was a problem hiding this comment.
Indeed, that's a trick I had from clef:
Line 308 in ccdbe29
Basically in embd mode, all tokens are outputs, so we expect to have one output row per input token. However, clef-based models produce one row per option (which is calculated over a span of tokens), so it's like many rows are "merged" into one final score output row, that's why other non-output need to be 0 (padding). This can be improved via a follow-up PR
CISC
approved these changes
Oct 12, 2026
| def filter_tensors(cls, item: tuple[str, Callable[[], Tensor]]) -> tuple[str, Callable[[], Tensor]] | None: | ||
| name, gen = item | ||
| # the backbone is a bare Qwen3Model, its tensors have no "model." prefix | ||
| return super().filter_tensors(("model." + name, gen)) |
Member
There was a problem hiding this comment.
I think it makes sense to add the sans-model. names instead, I already had it done for Qwen3.5 in 75118a3
ggerganov
approved these changes
Oct 12, 2026
CISC
added
the
highlight
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Add vllm decision2 model support: https://huggingface.co/collections/vllm-sr/decision-20
Requirements