AnkitAI/Parable-Granite-4.1-3B-Claude-Fable-5-GGUF

Remove 1 superseded/misnamed GGUF(s)

v3.1 quants: rebuilt from the promoted v3.1 weights (F16)

v3.1 quants: rebuilt from the promoted v3.1 weights (Q8_0)

v3.1 quants: rebuilt from the promoted v3.1 weights (Q6_K)

v3.1 quants: rebuilt from the promoted v3.1 weights (Q5_K_M)

v3.1 quants: rebuilt from the promoted v3.1 weights (Q4_K_M)

v3.1: souped seeds 42+43, merged at scale 0.6

v2.1: merge recalibrated to scale 0.6

Reframe evaluation guidance: same facts, buyer guidance instead of verdicts

Add measured BFCL V3 AST results (base vs Parable, identical harness)

Upload README.md with huggingface_hub

Plain link footer

Add maker link

Upload README.md with huggingface_hub

card: side-by-side output comparison, intelligence-density note, KV-cache tip

card: win-framed hero table, capability framing, closing CTA

card: single merged benchmark table, wins-first ordering

v2 card rework: headline-first, quant fit table, runbook style

v2: seed-averaged weights (F16)

v2 card: corpus v2, new recipe, updated benchmarks, cookbook usage

v2: seed-averaged weights, corpus v2 recipe (Q8_0)

v2: seed-averaged weights, corpus v2 recipe (Q6_K)

v2: seed-averaged weights, corpus v2 recipe (Q5_K_M)

v2: seed-averaged weights, corpus v2 recipe (Q4_K_M)

Add files using upload-large-folder tool

Upload folder using huggingface_hub

initial commit