Update README.md
Browse files
README.md
CHANGED
|
@@ -12,8 +12,12 @@ library_name: diffusers
|
|
| 12 |
<div align="center">
|
| 13 |
|
| 14 |
[](https://huggingface.co/Tongyi-MAI/Z-Image-Turbo) 
|
|
|
|
| 15 |
[](https://www.modelscope.cn/models/Tongyi-MAI/Z-Image-Turbo) 
|
| 16 |
-
[. It offers **⚡️sub-second inference latency⚡️** on enterprise-grade H800 GPUs and fits comfortably within **16G VRAM consumer devices**. It excels in photorealistic image generation, bilingual text rendering (English & Chinese), and robust instruction adherence.
|
| 29 |
|
|
|
|
|
|
|
| 30 |
- ✍️ **Z-Image-Edit** – A variant fine-tuned on Z-Image specifically for image editing tasks. It supports creative image-to-image generation with impressive instruction-following capabilities, allowing for precise edits based on natural language prompts.
|
| 31 |
|
| 32 |
### 📥 Model Zoo
|
| 33 |
|
| 34 |
-
| Model | Hugging Face
|
| 35 |
-
| :---
|
| 36 |
-
| **Z-Image-Turbo** | [ architecture. In this setup,
|
|
| 61 |
|
| 62 |
|
| 63 |
### 📈 Performance
|
| 64 |
-
According to the Elo-based Human Preference Evaluation (on [AI Arena](https://aiarena.alibaba-inc.com/corpora/arena/
|
| 65 |
|
| 66 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
|
| 68 |
### 🚀 Quick Start
|
| 69 |
|
| 70 |
```python
|
| 71 |
import torch
|
| 72 |
-
from diffusers import ZImagePipeline
|
| 73 |
|
| 74 |
# 1. Load the pipeline
|
| 75 |
# Use bfloat16 for optimal performance on supported GPUs
|
|
@@ -134,25 +145,17 @@ Our core insight behind DMDR is that Reinforcement Learning (RL) and Distributio
|
|
| 134 |
|
| 135 |
|
| 136 |
|
| 137 |
-
> More resources for DMDR will be released soon.
|
| 138 |
-
|
| 139 |
-
## ⏬ Download
|
| 140 |
-
|
| 141 |
-
```bash
|
| 142 |
-
hf download --model 'Tongyi-MAI/Z-Image-Turbo' --local_dir 'Z-Image-Turbo'
|
| 143 |
-
```
|
| 144 |
-
|
| 145 |
## 📜 Citation
|
| 146 |
|
| 147 |
If you find our work useful in your research, please consider citing:
|
| 148 |
|
| 149 |
```bibtex
|
| 150 |
@misc{z-image-2025,
|
| 151 |
-
title={Z-Image: An Efficient Image
|
| 152 |
-
author={
|
| 153 |
year={2025},
|
| 154 |
publisher={GitHub},
|
| 155 |
journal={GitHub repository},
|
| 156 |
howpublished={\url{https://github.com/Tongyi-MAI/Z-Image}}
|
| 157 |
}
|
| 158 |
-
```
|
|
|
|
| 12 |
<div align="center">
|
| 13 |
|
| 14 |
[](https://huggingface.co/Tongyi-MAI/Z-Image-Turbo) 
|
| 15 |
+
[](ttps://huggingface.co/spaces/Tongyi-MAI/Z-Image-Turbo) 
|
| 16 |
[](https://www.modelscope.cn/models/Tongyi-MAI/Z-Image-Turbo) 
|
| 17 |
+
[](https://www.modelscope.cn/aigc/imageGeneration?tab=advanced&versionId=469191&modelType=Checkpoint&sdVersion=Z_IMAGE_TURBO&modelUrl=modelscope%253A%252F%252FTongyi-MAI%252FZ-Image-Turbo%253Frevision%253Dmaster%7D%7BOnline) 
|
| 18 |
+
[](assets/Z-Image-Gallery.pdf) 
|
| 19 |
+
[](https://modelscope.cn/studios/Tongyi-MAI/Z-Image-Gallery/summary) 
|
| 20 |
+
|
| 21 |
|
| 22 |
Welcome to the official repository for the Z-Image（造相）project!
|
| 23 |
|
|
|
|
| 27 |
|
| 28 |
## ✨ Z-Image
|
| 29 |
|
| 30 |
+
Z-Image is a powerful and highly efficient image generation model with **6B** parameters. It is currently has three variants:
|
| 31 |
|
| 32 |
- 🚀 **Z-Image-Turbo** – A distilled version of Z-Image that matches or exceeds leading competitors with only **8 NFEs** (Number of Function Evaluations). It offers **⚡️sub-second inference latency⚡️** on enterprise-grade H800 GPUs and fits comfortably within **16G VRAM consumer devices**. It excels in photorealistic image generation, bilingual text rendering (English & Chinese), and robust instruction adherence.
|
| 33 |
|
| 34 |
+
- 🧱 **Z-Image-Base** – The non-distilled foundation model. By releasing this checkpoint, we aim to unlock the full potential for community-driven fine-tuning and custom development.
|
| 35 |
+
|
| 36 |
- ✍️ **Z-Image-Edit** – A variant fine-tuned on Z-Image specifically for image editing tasks. It supports creative image-to-image generation with impressive instruction-following capabilities, allowing for precise edits based on natural language prompts.
|
| 37 |
|
| 38 |
### 📥 Model Zoo
|
| 39 |
|
| 40 |
+
| Model | Hugging Face | ModelScope |
|
| 41 |
+
| :--- |:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
|
| 42 |
+
| **Z-Image-Turbo** | [](https://huggingface.co/Tongyi-MAI/Z-Image-Turbo) <br> [](https://huggingface.co/spaces/Tongyi-MAI/Z-Image-Turbo) | [](https://www.modelscope.cn/models/Tongyi-MAI/Z-Image-Turbo) <br> [](https://www.modelscope.cn/aigc/imageGeneration?tab=advanced&versionId=469191&modelType=Checkpoint&sdVersion=Z_IMAGE_TURBO&modelUrl=modelscope%3A%2F%2FTongyi-MAI%2FZ-Image-Turbo%3Frevision%3Dmaster) |
|
| 43 |
+
| **Z-Image-Base** | *To be released* | *To be released* |
|
| 44 |
+
| **Z-Image-Edit** | *To be released* | *To be released* |
|
| 45 |
|
| 46 |
### 🖼️ Showcase
|
| 47 |
|
|
|
|
| 67 |
|
| 68 |
|
| 69 |
### 📈 Performance
|
| 70 |
+
According to the Elo-based Human Preference Evaluation (on [AI Arena](https://aiarena.alibaba-inc.com/corpora/arena/leaderboard?arenaType=T2I)), Z-Image-Turbo shows highly competitive performance against other leading models, while achieving state-of-the-art results among open-source models.
|
| 71 |
|
| 72 |
+
<p align="center">
|
| 73 |
+
<a href="https://aiarena.alibaba-inc.com/corpora/arena/leaderboard?arenaType=T2I">
|
| 74 |
+
<img src="assets/leaderboard.webp" alt="Z-Image Elo Rating on AI Arena"/><br />
|
| 75 |
+
<span style="font-size:1.05em; cursor:pointer; text-decoration:underline;"> Click to view the full leaderboard</span>
|
| 76 |
+
</a>
|
| 77 |
+
</p>
|
| 78 |
|
| 79 |
### 🚀 Quick Start
|
| 80 |
|
| 81 |
```python
|
| 82 |
import torch
|
| 83 |
+
from diffusers import ZImagePipeline,
|
| 84 |
|
| 85 |
# 1. Load the pipeline
|
| 86 |
# Use bfloat16 for optimal performance on supported GPUs
|
|
|
|
| 145 |
|
| 146 |
|
| 147 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 148 |
## 📜 Citation
|
| 149 |
|
| 150 |
If you find our work useful in your research, please consider citing:
|
| 151 |
|
| 152 |
```bibtex
|
| 153 |
@misc{z-image-2025,
|
| 154 |
+
title={Z-Image: An Efficient Image Generation Foundation Model with Single-Stream Diffusion Transformer},
|
| 155 |
+
author={Tongyi Lab},
|
| 156 |
year={2025},
|
| 157 |
publisher={GitHub},
|
| 158 |
journal={GitHub repository},
|
| 159 |
howpublished={\url{https://github.com/Tongyi-MAI/Z-Image}}
|
| 160 |
}
|
| 161 |
+
```
|