Posts
All the articles I've posted.
Building llama.cpp with CUDA in an NVIDIA HPC SDK Environment
Configure CMake to enable the CUDA backend while selecting GCC/G++ for host code and NVCC for GPU code to avoid compiler-flag incompatibilities.
Building llama.cpp in an Environment Without curl Headers
Supply CMake with headers from matching curl source and the installed libcurl shared library to compile llama.cpp without installing curl development packages.
Getting Fill-In-the-Middle Autocomplete Working in VS Code Continue with llama.cpp
Assign a dedicated autocomplete model to llama-server in Continue's YAML configuration, tune suggestion latency, and keep chat endpoints separate when troubleshooting FIM completion.
Notes on Tabby: Llama.cpp, Model Caching, and Access Tokens
Notes on Tabby's local inference backend, configurable model cache directory, model registry, and browser-based account setup for obtaining an access token.