Files

README.de.md

README.es.md

docs [22]

examples [117]

demo-corpus [117]
- data [114]
  - research [114]
    daily [6]
    2026-05-05 [1]
    digest.md
    2026-04-30 [1]
    digest.md
    2026-04-25 [1]
    digest.md
    2026-04-20 [1]
    digest.md
    2026-04-15 [1]
    digest.md
    2026-04-10 [1]
    digest.md
    papers [90]
    arxiv-1607-02565 [2]
    abstract.md
    paper.md
    arxiv-1812-04605 [2]
    abstract.md
    paper.md
    arxiv-2002-10099 [2]
    abstract.md
    paper.md
    arxiv-2003-08934 [2]
    abstract.md
    paper.md
    arxiv-2008-02268 [2]
    abstract.md
    paper.md
    arxiv-2011-12948 [2]
    abstract.md
    paper.md
    arxiv-2103-13415 [2]
    abstract.md
    paper.md
    arxiv-2103-14024 [2]
    abstract.md
    paper.md
    arxiv-2104-06405 [2]
    abstract.md
    paper.md
    arxiv-2104-10078 [2]
    abstract.md
    paper.md
    arxiv-2105-06468 [2]
    abstract.md
    paper.md
    arxiv-2106-10689 [1]
    abstract.md
    arxiv-2106-12052 [2]
    abstract.md
    paper.md
    arxiv-2106-13228 [2]
    abstract.md
    paper.md
    arxiv-2108-10869 [2]
    abstract.md
    paper.md
    arxiv-2111-12077 [2]
    abstract.md
    paper.md
    arxiv-2112-05131 [1]
    abstract.md
    arxiv-2112-12130 [2]
    abstract.md
    paper.md
    arxiv-2201-05989 [1]
    abstract.md
    arxiv-2206-00665 [2]
    abstract.md
    paper.md
    arxiv-2209-14988 [1]
    abstract.md
    arxiv-2211-10440 [1]
    abstract.md
    arxiv-2301-10241 [2]
    abstract.md
    paper.md
    arxiv-2302-12249 [2]
    abstract.md
    paper.md
    arxiv-2303-11328 [2]
    abstract.md
    paper.md
    arxiv-2304-04278 [2]
    abstract.md
    paper.md
    arxiv-2304-14377 [2]
    abstract.md
    paper.md
    arxiv-2305-16213 [1]
    abstract.md
    arxiv-2306-17843 [2]
    abstract.md
    paper.md
    arxiv-2308-04079 [2]
    abstract.md
    paper.md
    arxiv-2309-16653 [1]
    abstract.md
    arxiv-2310-08528 [2]
    abstract.md
    paper.md
    arxiv-2311-04400 [2]
    abstract.md
    paper.md
    arxiv-2311-06214 [2]
    abstract.md
    paper.md
    arxiv-2311-11700 [2]
    abstract.md
    paper.md
    arxiv-2311-12775 [2]
    abstract.md
    paper.md
    arxiv-2311-13384 [2]
    abstract.md
    paper.md
    arxiv-2312-00109 [2]
    abstract.md
    paper.md
    arxiv-2312-02121 [2]
    abstract.md
    paper.md
    arxiv-2312-03203 [1]
    abstract.md
    arxiv-2312-06741 [2]
    abstract.md
    paper.md
    arxiv-2401-04099 [2]
    abstract.md
    paper.md
    arxiv-2402-03307 [2]
    abstract.md
    paper.md
    arxiv-2402-05054 [2]
    abstract.md
    paper.md
    arxiv-2402-07207 [2]
    abstract.md
    paper.md
    arxiv-2403-02151 [2]
    abstract.md
    paper.md
    arxiv-2403-14627 [1]
    abstract.md
    arxiv-2403-16292 [2]
    abstract.md
    paper.md
    arxiv-2403-17888 [2]
    abstract.md
    paper.md
    arxiv-2404-06109 [1]
    abstract.md
    questions [3]
    dynamic-reconstruction-scale.md
    inference-cost-vs-fidelity.md
    multi-view-consistency.md
    repos [13]
    ashawkey-stable-dreamfusion [1]
    about.md
    bmild-nerf [1]
    about.md
    cvlab-columbia-zero123 [1]
    about.md
    dreamgaussian-dreamgaussian [1]
    about.md
    graphdeco-inria-gaussian-splatting [2]
    about.md
    readme.md
    hustvl-4dgaussians [1]
    about.md
    nerfstudio-project-gsplat [1]
    about.md
    nvlabs-instant-ngp [1]
    about.md
    princeton-vl-droid-slam [1]
    about.md
    sxyu-svox2 [1]
    about.md
    totoro97-neus [1]
    about.md
    vast-ai-research-triposr [1]
    about.md
    weekly [2]
    2026-W18 [1]
    synthesis.md
    2026-W17 [1]
    synthesis.md
- INVENTORY.md
- LICENSES.md
- README.md

plugin [11]

commands [9]
README.md
skills [1]
- using-tesserae [1]
  - SKILL.md

prompts [1]

research_graph_extraction.md

tests [9]

fixtures [9]
- synthesis [1]
  - README.md
- wiki_corpus [8]
  - data [6]
    research [6]
    daily [5]
    2026-04-26 [2]
    …more
    …more
    2026-04-25 [3]
    …more
    …more
    weekly [1]
    2026-W17 [1]
    …more
  - docs [2]
    architecture.md
    quickstart.md

569 B · updated 2026-05-19 · md

paper.md

tests/fixtures/wiki_corpus/data/research/daily/2026-04-25/papers/2604.20329/paper.md

논문 분석: 2604.20329

arxiv: https://arxiv.org/abs/2604.20329
분석일: 2026-04-25

Image Generators are Generalist Vision Learners

저자: Valentin Gabeur, Songyou Peng, Kaiming He.

Vision Banana instruction-tunes Nano Banana Pro NBP on a mixture of image-generation data and small vision task data. By parameterising the output space of segmentation, metric depth estimation, and surface normal estimation as RGB images, perception is recast as image generation. The generalist model matches Segment Anything Model 3 and Depth Anything zero-shot.