v0.3.17-cu126-Basic-linux-20251202
·
9 commits
to main
since this release
update: Update Submodule vendor/llama.cpp 7f8ef50..746f9ee
feat: Sync llama.cpp API 20251202
feat: refactor: optimize LlamaGrammar class code
feat: Update llama_grammar.py from vendor/llama.cpp/examples/json-schema-to-grammar.py
feat: Enhanced text-based bootstrapping for Vulkan compilation and improved recognition of Win32 Vulkan libraries.
feat: Enhance Windows CUDA path detection logic
feat: Add new workflow with Basic options for Linux platform (CUDA 12.4, 12.6, 12.8)