Dflash speculator#24837
Draft
kashif wants to merge 15 commits into
Draft
Conversation
masking verified correct (80% accept, greedy-equivalent) but slower than the 472 t/s baseline until the per-layer cross-K/V cache lands - see dflash.md P2b.
…baseline) step-A fixed-shape+mask foundation preserved at 1a91884 for the P2b kv-cache (dflash.md).
raw prompt acceptance ~4% -> 80% (459 t/s) with LLAMA_SPEC_NO_THINK; matches PR/blog.
add a DFlash section + --spec-type entry to docs/speculative.md and the server/cli README option lists.
github-actions
Bot
added
documentation
Collaborator
|
wrong push target ? |
Contributor
|
I told him he could push in draft. I tested it at home, it's interesting. It looks clean. that would be too much all at once for 1 PR |
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
3 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Overview
POC DFlash speculator in llama.cpp using the new api vs the previous draft PR #22105
Additional information
Requirements