qwen4exp: support draft-head-only GGUFs (unsloth layout) + fix draft-load regression - #28097

qwen4exp: support draft-head-only GGUFs (unsloth layout) + fix draft-load regression#28097

TheArchitectit wants to merge 4 commits into

ggml-org:masterggml-org/llama.cpp:masterfrom

TheArchitectit:qwen4exp-draft-head-fixTheArchitectit/llama.cpp:qwen4exp-draft-head-fixCopy head branch name to clipboard