Tag
Demonstrates achieving 256k context on a 16GB RTX 4080 Super using Ternary Bonsai 27B Q2_0 model with llama.cpp, achieving up to 141 tok/s generation speed.