Skip to content

Commit e1f3ccb

Browse files
linmoumoulinmoumou
authored andcommitted
docs: expand Luma Uni-1 guide with architecture context and strengths
Add missing content from the official Luma Uni-1 launch blog post: - Architecture background: Uni-1 is a decoder-only autoregressive transformer (not a diffusion model), reasons before generating - Expanded 'What makes Uni-1 different' with text rendering, web search grounding, multilingual prompts, multi-panel output - New 'Strengths' section covering photorealism, illustration, old photo restoration, surreal compositions, text rendering, image editing, reference-guided generation, multi-panel output - New 'Getting started' section with Cloud template links and Desktop instructions - All changes mirrored in EN, ZH, and JA translations
1 parent 695821e commit e1f3ccb

3 files changed

Lines changed: 84 additions & 6 deletions

File tree

ja/tutorials/partner-nodes/luma/luma-uni-1.mdx

Lines changed: 28 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,9 @@ description: "Luma の統一モデル Uni-1 による画像生成と精密編集
44
sidebarTitle: "Luma Uni-1"
55
---
66

7-
Uni-1 は、画像生成と精密編集の両方に対応する Luma 初の統一モデルです。出力品質だけでなく、**コントロール性**こそがその真価です。新しい画像をゼロから生成することも、既存の画像を外科手術のように精密に編集することもでき、参照画像・シード・構造化されたプロンプトを使って結果を意図通りに導けます。
7+
Uni-1 は、画像生成と精密編集の両方に対応する Luma 初の統一モデルです。ほとんどの画像モデルとは異なり、Uni-1 は拡散モデルではなく、**デコーダーのみの自己回帰型トランスフォーマー**です。テキストと画像をひとつのインターリーブされたシーケンスとして扱い、時間・空間・論理を単一のアーキテクチャで統合的にモデリングします。その結果、モデルは生成前にプロンプトを**推論**し、指示を分解し、制約を解決し、最先端の LLM のように構図を計画します。まず考え、それから描きます。
8+
9+
出力品質だけでなく、**コントロール性**こそがその真価です。新しい画像をゼロから生成することも、既存の画像を外科手術のように精密に編集することもでき、参照画像・シード・構造化されたプロンプトを使って結果を意図通りに導けます。
810

911
本ガイドでは、最初の 1 枚から高度なマルチリファレンス構成まで、実用的なメンタルモデルとワークフローを紹介します。
1012

@@ -15,9 +17,26 @@ Uni-1 は、画像生成と精密編集の両方に対応する Luma 初の統
1517
- ニッチで具体的なビジュアルスタイルへの**強力なコントロール**
1618
- 再現性と制御された反復のための**シード対応**
1719
- 縦長から横長まで**9 種類のアスペクト比**
20+
- **テキストレンダリング** — 画像内で読める文字、インフォグラフィックやポスターに最適
21+
- **ウェブ検索グラウンディング** — 実世界のコンテキストを活用
22+
- **多言語プロンプト**対応
23+
- **マルチパネル出力** — 複数フレーム間でのキャラクター・シーン一貫性
1824

1925
モードと参照画像の制御方法を理解すれば、Uni-1 を理解したと言えます。
2026

27+
## 得意分野
28+
29+
Uni-1 は幅広いタスクで優れた性能を発揮します:
30+
31+
- **フォトリアリズム**と素材の正確な再現
32+
- **イラストとスタイライズドアート**、強い美的コントロール
33+
- **古写真の修復**とビンテージ再現
34+
- **シュールでコンセプチュアル**な構図
35+
- **テキストレンダリング** — 画像内で読める文字、インフォグラフィックやポスターに最適
36+
- **画像編集**とマルチターンでのリファイン
37+
- **参照画像ガイド生成**、人物のアイデンティティ保持
38+
- **マルチパネル出力** — 複数フレーム間でのキャラクター・シーン一貫性
39+
2140
## 核となる区別:Create か Modify か
2241

2342
Uni-1 のすべては、シンプルな問いから始まります:**新しいものを作るのか、それとも既にあるものを変えるのか?**
@@ -32,7 +51,14 @@ Uni-1 のすべては、シンプルな問いから始まります:**新しい
3251
- 出力が入力画像の**バリエーション**であるべきなら Modify
3352
- 入力に**着想を得つつ新しい**ものなら Create
3453

35-
## 最初の 1 枚
54+
## 始め方
55+
56+
Uni-1 を使用する方法は 2 つあります:
57+
58+
- **Comfy Cloud**:テンプレートパネルから [Create](https://cloud.comfy.org/?template=api_luma_uni_1_image_create) または [Modify](https://cloud.comfy.org/?template=api_luma_uni_1_image_edit) テンプレートを読み込む。
59+
- **デスクトップ版**:ComfyUI を最新バージョンにアップデートし、ノードライブラリから **Luma UNI-1 Image** ノードを見つけてワークフローに接続。
60+
61+
### 最初の 1 枚
3662

3763
1. **Create Image > Uni-1** を開く
3864
2. プロンプトを書く

tutorials/partner-nodes/luma/luma-uni-1.mdx

Lines changed: 28 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,9 @@ description: "A practical guide to Luma's Uni-1 unified model for image creation
44
sidebarTitle: "Luma Uni-1"
55
---
66

7-
Uni-1 is Luma's first unified model designed for both image creation and precision editing. Beyond raw output quality, what makes it powerful is **control**: you can generate entirely new images or surgically modify existing ones, while guiding the system with references, seeds, and structured prompts.
7+
Uni-1 is Luma's first unified model designed for both image creation and precision editing. Unlike most image models, Uni-1 isn't a diffusion model — it's a **decoder-only autoregressive transformer** that treats text and images as a single interleaved sequence, jointly modeling time, space, and logic in one architecture. The result: a model that reasons about your prompt before generating — decomposing instructions, resolving constraints, and planning composition like a frontier LLM. It thinks first, then draws.
8+
9+
Beyond raw output quality, what makes it powerful is **control**: you can generate entirely new images or surgically modify existing ones, while guiding the system with references, seeds, and structured prompts.
810

911
This guide gives you a working mental model and practical workflows, from your first image to advanced multi-reference setups.
1012

@@ -15,9 +17,26 @@ This guide gives you a working mental model and practical workflows, from your f
1517
- **Strong control over niche and specific visual styles**
1618
- **Seed support** for reproducibility and controlled iteration
1719
- **Nine aspect ratios**, from ultra-tall to ultra-wide
20+
- **Text rendering** that's actually readable
21+
- **Web search grounding** for real-world context
22+
- **Multilingual prompts**
23+
- **Multi-panel output** with temporal consistency
1824

1925
If you understand the modes and how to control references, you understand Uni-1.
2026

27+
## Strengths
28+
29+
Uni-1 excels across a wide range of tasks:
30+
31+
- **Photorealism** with material accuracy
32+
- **Illustration & stylized art** with strong aesthetic control
33+
- **Old photo restoration** and vintage reproduction
34+
- **Surreal and conceptual** compositions
35+
- **Text rendering** — readable text inside images, great for infographics and posters
36+
- **Image editing** and multi-turn refinement
37+
- **Reference-guided generation** with identity preservation
38+
- **Multi-panel output** — consistent characters/scenes across multiple frames
39+
2140
## The core distinction: Create vs Modify
2241

2342
Everything in Uni-1 starts with one question: **Am I creating something new, or changing something that already exists?**
@@ -32,7 +51,14 @@ When in doubt:
3251
- If the output should look like a **version of your input**, use Modify
3352
- If it should feel **inspired but new**, use Create
3453

35-
## Your first image
54+
## Getting started
55+
56+
You can use Uni-1 in two ways:
57+
58+
- **Comfy Cloud**: Load the [Create](https://cloud.comfy.org/?template=api_luma_uni_1_image_create) or [Modify](https://cloud.comfy.org/?template=api_luma_uni_1_image_edit) template from the Templates panel.
59+
- **Desktop**: Update ComfyUI to the latest version, find the **Luma UNI-1 Image** node via the Node Library, and connect it to your workflow.
60+
61+
### Your first image
3662

3763
1. Open **Create Image > Uni-1**
3864
2. Write a prompt

zh/tutorials/partner-nodes/luma/luma-uni-1.mdx

Lines changed: 28 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,9 @@ description: "Luma Uni-1 统一模型使用指南:图像创建与精准编辑"
44
sidebarTitle: "Luma Uni-1"
55
---
66

7-
Uni-1 是 Luma 推出的首款统一模型,同时支持图像创建(Create)与精准编辑(Modify)。它真正的优势不仅在于输出质量,更在于**可控性**——你既可以从零生成全新图像,也可以对已有图像进行精细编辑,并通过参考图、种子(seed)和结构化提示词来引导生成结果。
7+
Uni-1 是 Luma 推出的首款统一模型,同时支持图像创建(Create)与精准编辑(Modify)。与大多数图像模型不同,Uni-1 不是扩散模型——它是一个 **decoder-only autoregressive transformer(仅解码器自回归变换器)**,将文本和图像视为单一交错序列,在单一架构中联合建模时间、空间和逻辑。结果是:模型会在生成之前**先推理你的提示词**——分解指令、解决约束条件、像一个前沿 LLM 一样规划构图。先思考,再绘画。
8+
9+
它真正的优势不仅在于输出质量,更在于**可控性**——你既可以从零生成全新图像,也可以对已有图像进行精细编辑,并通过参考图、种子(seed)和结构化提示词来引导生成结果。
810

911
本指南将帮助你建立可用的心智模型与实用工作流,从你的第一张图像到进阶的多参考图组合。
1012

@@ -15,9 +17,26 @@ Uni-1 是 Luma 推出的首款统一模型,同时支持图像创建(Create
1517
- **强力控制**特定的小众视觉风格
1618
- **支持种子(seed)**,便于复现和受控迭代
1719
- **9 种宽高比**,从超高到超宽
20+
- **文字渲染** — 图像内可读的文字,非常适合信息图和海报
21+
- **网页搜索**作为背景知识来源
22+
- **多语言提示词**支持
23+
- **多面板输出** — 跨多帧保持角色/场景一致性
1824

1925
理解了模式与参考图控制方式,你就理解了 Uni-1。
2026

27+
## 强项
28+
29+
Uni-1 在广泛的任务中表现优异:
30+
31+
- **照片级真实感**与材质精度
32+
- **插画与风格化艺术**,审美控制力强
33+
- **老照片修复**与复古再现
34+
- **超现实与概念性**构图
35+
- **文字渲染** — 图像内可读的文字,擅长信息图与海报
36+
- **图像编辑**与多轮细化
37+
- **参考图引导生成**,可保持人物身份一致性
38+
- **多面板输出** — 多帧间保持角色/场景一致
39+
2140
## 核心区别:Create 还是 Modify
2241

2342
使用 Uni-1 之前,先问自己一个问题:**我是要创建全新的内容,还是要修改已有的内容?**
@@ -32,7 +51,14 @@ Uni-1 是 Luma 推出的首款统一模型,同时支持图像创建(Create
3251
- 如果输出应当像输入图像的**某个版本**,使用 Modify
3352
- 如果输出应当**受其启发但全新**,使用 Create
3453

35-
## 你的第一张图像
54+
## 开始使用
55+
56+
你可以通过以下两种方式使用 Uni-1:
57+
58+
- **Comfy Cloud**:从模板面板加载 [Create](https://cloud.comfy.org/?template=api_luma_uni_1_image_create)[Modify](https://cloud.comfy.org/?template=api_luma_uni_1_image_edit) 模板。
59+
- **桌面版**:将 ComfyUI 更新到最新版,通过节点库找到 **Luma UNI-1 Image** 节点并连接到你的工作流中。
60+
61+
### 你的第一张图像
3662

3763
1. 打开 **Create Image > Uni-1**
3864
2. 编写提示词

0 commit comments

Comments
 (0)