esatapedico/Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP-NVFP4-GGUF

Upload README.md with huggingface_hub

Update model card: no longer recommend DRY anti-loop sampling

MEDIUM deployed context 307200 (training ctx 262144 documented)

Note final deployed contexts (LOW 307200 / MEDIUM 262144 / HIGH 221184 / VERY-HIGH 204800)

Add family naive observations from 180k single-run tests

DRY anti-loop sampler applies to all four tiers incl. LOW

Clean up card: remove em/en dashes

Add description to model card frontmatter

Update model card to family wording (4 tiers)

Add per-tier override maps

Add MEDIUM/HIGH/VERY-HIGH tiers; refresh LOW with metadata

Update README: full model name with MTP marker

Remove stale Qwen3.6-27B-Fable-Fusion-711-NVFP4-LOW.gguf

Rename GGUF to Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP-NVFP4-LOW.gguf

Update vision note: any DavidAU mmproj variant works (BF16/F16/F32)

Update model card: add links to base models, remove YAML reference

Remove personal LocalAI config (not for public distribution)

Add NVFP4-LOW hybrid GGUF (15.49 GB)

Add overrides.txt

Add Qwen3.6-27B-Fable-Fusion-711-MTP-maci0-NVFP4-LOW.yaml

Add README.md

initial commit