# HVision-NKU/StoryDiffusion

Accepted as [NeurIPS 2024] Spotlight Presentation Paper

Repository: https://github.com/HVision-NKU/StoryDiffusion
Canonical: https://ross.abutalabs.com/products/storydiffusion
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Last push: 2024-09-26T02:17:52+00:00

## Health v2 (maintenance only)
Score: 24/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 61
- inputs: {"age_days": 864, "days_push": 707, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 6452, forks 641 (observed 2026-08-28T04:09:43.510081+00:00)

## What it is
StoryDiffusion is the official implementation of a NeurIPS 2024 Spotlight paper introducing Consistent Self-Attention for character-consistent long-range image generation and a motion predictor for long video generation. It plugs into SD1.5 and SDXL-based diffusion models to generate consistent image sequences, comics, and videos from text prompts.

## Use cases
- generate consistent comic strips from text prompts
- create character-consistent image sequences with stable diffusion
- generate long videos from a sequence of images
- make a visual story with the same character across many images
- add consistent self-attention to my SDXL pipeline
- turn text prompts into an illustrated storybook

## When to choose
- you need the same character to appear consistently across many generated images
- you want to build comics or visual stories from text prompts
- you want to extend consistent images into long videos
- you already use SD1.5 or SDXL pipelines and want a hot-pluggable consistency module

## When to avoid
- you need single unrelated images with no character consistency
- you lack a GPU or the compute for diffusion model inference
- you need production-grade video generation rather than a research implementation
- you need fine-grained control over video motion beyond the motion predictor

## Facets
- artifact type: library
- maturity: active
- function: image-processing, video-processing, machine-learning, deep-learning
- domain: artificial-intelligence, image-processing, deep-learning
- platform: python
- tags: diffusion-models, stable-diffusion, text-to-image, image-to-video, character-consistency, story-generation, comics, neurips-2024, video, gpu

## Member repositories
- HVision-NKU/StoryDiffusion (main) score 24

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:43.510081+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:44:39.132057+00:00, confidence not recorded.
  - readme: https://github.com/HVision-NKU/StoryDiffusion (fetched 2026-08-28T04:09:43.510081+00:00, sha 48d2ebf0ff9a)
- Data as of 2026-08-30T08:39:29.467469+00:00.