Modelpublic

Veetance/Narrative-Brain-5B-NF4

sourceHugging Faceapache-2.0updated 5mo agoView on Hugging Face

0likes22downloads

Model Card

Narrative-Brain-9B-NF4

This repository contains the local Narrative Brain runtime used by the Asset Editor Stealth refine path.

Model Details

—Runtime name: Narrative Brain 9B
—Actual checkpoint class: compact Qwen2-family causal LM
—Quantization: NF4 via bitsandbytes
—Primary use: prompt refinement, prompt enhancement, and in-house narrative steering inside the Asset Editor stack
—License: Apache-2.0

Runtime Notes

—The repository preserves the historical narrative_brain_9b path name used by the app.
—The current checkpoint is a compact runtime, not a full 9B-class Qwen block.
—Local logs identify the repaired architecture as Qwen2ForCausalLM after removal of an incorrect Qwen3 wrapper.

Included Runtime Components

—model.safetensors
—config.json
—generation_config.json
—chat_template.jinja
—special_tokens_map.json
—vocab.json
—merges.txt
—tokenizer.json
—tokenizer_config.json

Integration

Optimized for the Veetance Asset Editor Stealth refine runtime.

Links

—Qwen family reference: Qwen/Qwen2.5-1.5B-Instruct
—Core engine: Veetance Asset Editor
—Triage and development: Asset Editor GitHub

This repository mirrors the working local runtime bundle under models/narrative_brain_9b.