naklitechie/Qwen3.8-27B-DFlash2-ternary-bonsai2

Card: prompt-lookup + DFlash 2 stacking results and flag

Card: NVIDIA L4 / llama.cpp benchmark, llama.cpp usage, #261 pointer

Model card: browser use (LocalMind WebGPU, 1.18x on code)

model card: list the Q4_K_M GGUF

Q4_K_M GGUF of the round-3 drafter (dflash arch; for llama.cpp draft-dflash and the WebGPU engine)

model card

round-3 drafter, bf16

add config (from z-lab/Qwen3.8-27B-DFlash2)

initial commit