// image · deepseek-ai/Janus

Janus

Janus é um projeto de imagem de código aberto de deepseek-ai (deepseek-ai/Janus no GitHub). Tem 17.754 estrelas no GitHub e 2.232 forks, é publicado sob a licença MIT e foi atualizado pela última vez no GitHub em 1 de fevereiro de 2025.

DeepSeek's Janus series - unified multimodal models that both understand and generate images

★ 17,8 mil estrelas⑂ 2,2 mil forksMITatualizado fev 2025

Ver no GitHub ↗

▦ visual: o README deste projeto não inclui nenhuma captura incorporável,
então aqui vai um marcador. O guia abaixo está completo mesmo assim.

// dados do projeto

Dados do projeto

Repositório: deepseek-ai/Janus
Estrelas no GitHub: 17.754
Forks: 2.232
Licença: MIT
Última atualização no GitHub: 1 de fevereiro de 2025
Categoria: Imagem
Tipo de ficha: Guia de instalação completo

// como funciona

Como funciona

Choose a model size
Pick a Janus variant and size (for example Janus-Pro-7B or the 1.3B models) that fits your GPU.
Clone the repository
Get the code, which includes inference examples and demos for each model in the series.
Install dependencies
Create a Python 3.8+ environment and install the package in editable mode.
Run your first inference
Load the weights from Hugging Face and run multimodal understanding or text-to-image generation.

// recursos

Por que as pessoas gostam tanto

Unified understanding and generation

One model answers questions about images and generates images from text.

Decoupled visual encoding

Separate visual pathways for understanding and generation avoid the conflict that limited earlier unified models.

Janus-Pro improvements

An optimized training strategy, expanded data and larger scale improve instruction following and generation stability.

JanusFlow variant

Integrates rectified flow generation within the LLM framework for efficient unified modeling.

Gradio demos

Local Gradio apps (and hosted Hugging Face Spaces demos) for interactive testing.

FastAPI server

A FastAPI demo exposes the same functions as an API server.

// execute você mesmo

Execute você mesmo

Clone the repository
Get the source code from GitHub and enter the project directory.
```
git clone https://github.com/deepseek-ai/Janus.git
cd Janus
```
Install the package
Install Janus and its dependencies into your Python environment.
```
pip install -e .
```
Run a demo
Launch the Janus-Pro Gradio demo, or the FastAPI demo for API access.
```
python demo/app_januspro.py
# API variant: python demo/fastapi_app.py
```

// casos de uso

O que você pode criar

Build an assistant that can both describe a product photo and generate variations of it
Research unified multimodal architectures against task-specific models
Self-host a text-to-image generator with image-understanding in the same service
Evaluate Janus-Pro on your own vision-language benchmark set

// perguntas frequentes

Perguntas frequentes

Who maintains this project?

It is maintained by DeepSeek AI as an open-source research release on GitHub.

Is it free to use?

The code is MIT licensed; the model weights are under DeepSeek's model agreement (LICENSE-MODEL in the repo), so read that before commercial use.

Does it need a paid API key?

No API key or fee: models run locally on your GPU after a free Hugging Face download.

How does it compare to alternatives?

Compared with separate specialist models (a VLM plus an image generator), Janus unifies both in one model; compared with hosted APIs, it is fully self-hosted but needs your own GPU.

How do I get started?

Os outros projetos de Imagem com mais estrelas no catálogo do LibreAI — a lista realista de alternativas se Janus não for o ideal para você.

Stable Diffusion Webui★ 165,2 mil
Stable Diffusion web UI
ComfyUI★ 136,0 mil
The node-graph powerhouse for precise, professional AI image and video generation.
Immich★ 115,5 mil
High performance self-hosted photo and video management solution
PhotoCraft★ 73,2 mil
PhotoCraft — a clean-room Photoshop reimplementation in pure Rust: layers, masks, adjustment layers, type, vectors, brushes and real PSD files, offline, MIT / Apache-2.0
Fooocus★ 53,3 mil
Focus on prompting and generating

// mais em image

Mais em Imagem

Projetos de Imagem próximos no catálogo, os mais próximos em estrelas de Janus: continue explorando dos dois lados dele.

ml-stable-diffusion★ 18,0 mil
Stable Diffusion with Core ML on Apple Silicon
stable-diffusion-webui-colab★ 15,9 mil
stable diffusion webui colab
CodeFormer★ 18,2 mil
[NeurIPS 2022] Towards Robust Blind Face Restoration with Codebook Lookup Transformer
awesome-nano-banana-pro-prompts★ 13,5 mil
🍌 World's largest Nano Banana Pro prompt library — 10,000+ curated prompts with preview images, 16 languages. Google Gemini AI image generation. Free & open …
SAM 2★ 20,1 mil
SAM 2 — Meta's Segment Anything Model 2: promptable segmentation for images and video, Apache-2.0
darktable★ 13,2 mil
darktable is an open source photography workflow application and raw developer

Baixe o Janus

Grátis e aberto: pegue o código-fonte, dê uma estrela no repositório e comece a criar.

Ver no GitHub ↗← Mais ferramentas de Imagem