Repository navigation
Build on official llama.cpp; vendor the decision engine - #6
Merged
Merged
Conversation
The parallel-decision fork changed nothing in libllama: its engine runs on
the stock API. native/decision-engine.{h,cpp} is that engine (the class
lev_decision.cpp drives, without the fork's schema compiler and server
glue), with the two common/ batch helpers upstream dropped kept locally.
llama.cpp is pinned at release b11630, 589 commits past the fork's base,
until ggml-org/llama.cpp#29752 lands.
It read LEV_CHECKPOINTS or ../laya and ignored config.edn's :checkpoints-home, which jolt prepare honours.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The parallel-decision fork changed nothing in libllama: its decision engine runs on the stock API. So native/decision-engine.{h,cpp} now carries the engine, trimmed to the class lev_decision.cpp drives (the fork's schema compiler and server glue are left out), and llama.cpp is pinned at official release b11630, 589 commits past the fork's base. Upstream dropped the two common/ batch helpers the engine used; they are kept locally. The engine stays vendored until ggml-org/llama.cpp#29752 lands.
Measured back to back against the fork's library: the answers don't move (authored144 108/144, ECE 0.174 on both; llm-test, think-test and clef-test pass against their goldens), and upstream is 3-15% faster on every Jev workflow row. Numbers are in bench/README.md.
The release workflow, run on this branch with jolt nightly (JOLT_PIN=vnightly, which has jolt-lang/jolt#1341 and #1343), builds and self-tests the binary on mac and linux: run 38675927682.
Also: prepare_test now finds the checkpoints the way jolt prepare does, so config.edn's :checkpoints-home counts, not just LEV_CHECKPOINTS or ../laya.