plunderstruck/Qwen3.6-27B-OBLITERATED-MTP-ROCmFP4-GGUF
Repoint build instructions to charlie12345/ROCmFPX (ROCmFPX FP3/4/6/8 repo)
Upload README.md with huggingface_hub
consolidate to single best-balance rocmfp4 quant
consolidate to single best-balance rocmfp4 quant
Card: point launch -m at best-of-both (copy-paste vision works)
Card: best-of-both recommended + KL findings + bundled mmproj/vision flags
Add mmproj-F32.gguf (best-of-both / vision)
Add best-of-both recipe (all-dual body + f16 emb + Q6 head + imatrix)
Add Qwen ASCII logo header (centered logo+title, full-width spec table)
Correct mmproj/cache claim: source says only --cache-reuse is disabled w/ multimodal (not 'cache reuse entirely'); rest unmeasured. (27B: + upstream-benchmark pointer)
Claim audit: remove cross-model generalizations (Q6-head %, f16-draft-head, decode comparisons) — state per-model config only, no borrowed sibling measurements
Card: datasheet/neo-brutalist restyle (monospace, hard borders, spec grid, numbered sections); claims audited (f16 KV stated as config only)
Card: add Vision section (--mmproj + --image-min-tokens 1024) and bundled chat_template.jinja note
Add froggeric v20 fixed chat template (vision + inline think-toggle + tool anti-stall); use via --chat-template-file
card: add surfaced Prompt format & sampling block
final card: measured KL/PPL A/B + MTP acceptance (2.10x, 67.7%)
imatrix quant, Q6_K output head
imatrix quant, base (4-bit head)
WIP card: ROCmFP4 quant coming soon (build in progress)
initial commit