v1.0.1: Fix tokenizer loading, prompt formatting, cache wiring, MCP intelligence by devwhodevs · Pull Request #11 · devwhodevs/engraph

fix(llm): add shimmytok fallback for GGUF-embedded tokenizers

devwhodevs

committed

b478bec

fix(llm): apply prompt format in CandleEmbed embed_one and embed_batch

devwhodevs

committed

bd87d71

fix(store): clear FTS on reindex and use stored dim for vec table init

devwhodevs

committed

7c6ba89

fix(search): wire LLM cache into search_with_intelligence

devwhodevs

committed

9017388

fix(serve): wire orchestrator and reranker into MCP search handler

devwhodevs

committed

6a55e31

feat(llm): add BERT GGUF architecture support, switch default to all-MiniLM-L6-v2

devwhodevs

committed

dd21f41

feat: add accelerate feature flag for optimized CPU on macOS

devwhodevs

committed

7171792

fix: add indexing progress output, fix Qwen3 GGUF filename case

devwhodevs

committed

4892309

fix: use float32 RmsNorm for Metal GPU compatibility in Gemma embedding

devwhodevs

committed

20be487

refactor(llm): replace candle backend with llama-cpp-2 for Metal GPU support

devwhodevs

committed

ebb814b

feat(llm): switch to llama.cpp backend, fix embedding params

devwhodevs

committed

3db1ae5

style: cargo fmt

devwhodevs

committed

8c4f192

ci: install CMake on Ubuntu for llama.cpp build

devwhodevs

committed

f55dd85

feat(search): wire intelligence models into CLI search path

devwhodevs

committed

55f67ad

fix: singleton LlamaBackend and built-in tokenizer for orchestrator/reranker

devwhodevs

committed

5c7f285

fix(llm): global backend singleton, built-in tokenizers, wire CLI intelligence

devwhodevs

committed

8b85976

docs: update README, CHANGELOG, CLAUDE.md for llama.cpp backend

devwhodevs

committed

e8c159f

Navigation Menu

Repository navigation

v1.0.1: Fix tokenizer loading, prompt formatting, cache wiring, MCP intelligence - #11

v1.0.1: Fix tokenizer loading, prompt formatting, cache wiring, MCP intelligence#11
devwhodevs merged 17 commits into
mainfrom
hotfix/v1.0.1-tokenizer-and-review-fixes

Commits on Mar 25, 2026

fix(llm): add shimmytok fallback for GGUF-embedded tokenizers

fix(llm): apply prompt format in CandleEmbed embed_one and embed_batch

fix(store): clear FTS on reindex and use stored dim for vec table init

fix(search): wire LLM cache into search_with_intelligence

fix(serve): wire orchestrator and reranker into MCP search handler

feat(llm): add BERT GGUF architecture support, switch default to all-MiniLM-L6-v2

feat: add accelerate feature flag for optimized CPU on macOS

fix: add indexing progress output, fix Qwen3 GGUF filename case

fix: use float32 RmsNorm for Metal GPU compatibility in Gemma embedding

refactor(llm): replace candle backend with llama-cpp-2 for Metal GPU support

feat(llm): switch to llama.cpp backend, fix embedding params

style: cargo fmt

ci: install CMake on Ubuntu for llama.cpp build

feat(search): wire intelligence models into CLI search path

fix: singleton LlamaBackend and built-in tokenizer for orchestrator/reranker

fix(llm): global backend singleton, built-in tokenizers, wire CLI intelligence

docs: update README, CHANGELOG, CLAUDE.md for llama.cpp backend