Add qwen3.5-9b.mq4-awq-gptq-f2-lmhead-a100.hfq (AWQ+GPTQ MQ4 with AWQ-on-lm_head, A100-built)

Add Qwen3.5-9B MQ4 AWQ-GPTQ Q8conv F2 (CUDA path, alpha=0.55)

Add qwen3.5-9b.mq4-lloyd (Phase B WMMA prefill research quant, PR #197)

Add qwen3.5-9b.mq3-lloyd

Initial: research-gated MQ3-Lloyd + MQ4-Lloyd disclaimer

initial commit