<!-- ## **HunyuanDiT** -->

<p align="center">
  <img src="./asset/logo.png"  height=100>
</p>

# Hunyuan-DiT : A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding

-----

This repo contains PyTorch model definitions, pre-trained weights and inference/sampling code for our paper exploring Hunyuan-DiT. You can find more visualizations on our [project page](https://dit.hunyuan.tencent.com/).

> [**Hunyuan-DiT : A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding**](https://github.com/Tencent/HunyuanDiT/blob/main/asset/Hunyuan_DiT_Tech_Report_05140553.pdf) <br>
> Zhimin Li*, Jianwei Zhang*, Qin Lin, Jiangfeng Xiong, Yanxin Long, Xinchi Deng, Yingfang Zhang, Xingchao Liu, Minbin Huang, Zedong Xiao, Dayou Chen, Jiajun He, Jiahao Li, Wenyue Li, Chen Zhang, Rongwei Quan, Jianxiang Lu, Jiabin Huang, Xiaoyan Yuan, Xiaoxiao Zheng, Yixuan Li, Jihong Zhang, Chao Zhang, Meng Chen, Jie Liu, Zheng Fang, Weiyan Wang, Jinbao Xue, Yangyu Tao, JianChen Zhu, Kai Liu, Sihuan Lin, Yifu Sun, Yun Li, Dongdong Wang, Zhichao Hu, Xiao Xiao, Yan Chen, Yuhong Liu, Wei Liu, Di Wang, Yong Yang, Jie Jiang, Qinglin Lu‡
> <br>Tencent Hunyuan<br>

> [**DialogGen:Multi-modal Interactive Dialogue System for Multi-turn Text-to-Image Generation**](https://hunyuan-dialoggen.github.io/)<br>
> Minbin Huang*, Yanxin Long*, Xinchi Deng,  Ruihang Chu, Jiangfeng Xiong, Xiaodan Liang, Hong Cheng, Qinglin Lu&#8224;, Wei Liu
> <br>Chinese University of Hong Kong, Tencent Hunyuan, Shenzhen Campus of Sun Yat-sen University<br>

---

## 📑 Open-source Plan

- Hunyuan-DiT (Text-to-Image Model)
  - [x] Inference ✅
  - [x] Checkpoints ✅
  - [ ] Distillation Version (Coming soon ⏩️)
  - [ ] TensorRT Version (Coming soon ⏩️)
  - [ ] Training (Coming later ⏩️)
- DialogGen (Prompt Enhancement Model)
  - [x] Inference ✅
- [X] Web Demo (Gradio) ✅
- [X] Cli Demo ✅

## Contents
- [Hunyuan-DiT](#hunyuan-dit-a-powerful-multi-resolution-diffusion-transformer-with-fine-grained-chinese-understanding)
  - [Abstract](#abstract)
  - [🎉 Hunyuan-DiT Key Features](#hunyuan-dit-key-features)
    - [Chinese-English Bilingual DiT Architecture](#chinese-english-bilingual-dit-architecture)
    - [Multi-turn Text2Image Generation](#multi-turn-text2image-generation)
  - [📈 Comparisons](#comparisons)
  - [🎥 Visualization](#visualization)
  - [📜 Requirements](#requirements)
  - [🛠 Dependencies and Installation](#dependencies-and-installation)
  - [🧱 Download Pretrained Models](#download-pretrained-models)
  - [🔑 Inference](#inference)
    - [Using Gradio](#using-gradio)
    - [Using Command Line](#using-command-line)
    - [More Configurations](#more-configurations)
  - [🔗 BibTeX](#bibtex)

## **Abstract**

We present Hunyuan-DiT, a text-to-image diffusion transformer with fine-grained understanding of both English and Chinese. To construct Hunyuan-DiT, we carefully designed the transformer structure, text encoder, and positional encoding. We also build from scratch a whole data pipeline to update and evaluate data for iterative model optimization. For fine-grained language understanding, we train a Multimodal Large Language Model to refine the captions of the images. Finally, Hunyuan-DiT can perform multi-round multi-modal dialogue with users, generating and refining images according to the context.
Through our carefully designed holistic human evaluation protocol with more than 50 professional human evaluators, Hunyuan-DiT sets a new state-of-the-art in Chinese-to-image generation compared with other open-source models.


## 🎉 **Hunyuan-DiT Key Features**
### **Chinese-English Bilingual DiT Architecture**
Hunyuan-DiT is a diffusion model in the latent space, as depicted in figure below. Following the Latent Diffusion Model, we use a pre-trained Variational Autoencoder (VAE) to compress the images into low-dimensional latent spaces and train a diffusion model to learn the data distribution with diffusion models. Our diffusion model is parameterized with a transformer. To encode the text prompts, we leverage a combination of pre-trained bilingual (English and Chinese) CLIP and multilingual T5 encoder.
<p align="center">
  <img src="./asset/framework.png"  height=500>
</p>

### Multi-turn Text2Image Generation
Understanding natural language instructions and performing multi-turn interaction with users are important for a
text-to-image system. It can help build a dynamic and iterative creation process that bring the user’s idea into reality
step by step. In this section, we will detail how we empower Hunyuan-DiT with the ability to perform multi-round
conversations and image generation. We train MLLM to understand the multi-round user dialogue
and output the new text prompt for image generation.
<p align="center">
  <img src="./asset/mllm.png"  height=300>
</p>

## Comparisons
In order to comprehensively compare the generation capabilities of HunyuanDiT and other models, we constructed a 4-dimensional test set, including Text-Image Consistency, Excluding AI Artifacts, Subject Clarity, Aesthetic. More than 50 professional evaluators performs the evaluation.

<p align="center">
<table> 
<thead> 
<tr> 
    <th rowspan="2">Model</th> <th rowspan="2">Open Source</th> <th>Text-Image Consistency (%)</th> <th>Excluding AI Artifacts (%)</th> <th>Subject Clarity (%)</th> <th rowspan="2">Aesthetics (%)</th> <th rowspan="2">Overall (%)</th> 
</tr> 
</thead> 
<tbody> 
<tr> 
    <td>SDXL</td> <td>&#10004</td> <td>64.3</td> <td>60.6</td> <td>91.1</td> <td>76.3</td> <td>42.7</td> 
</tr> 
<tr> 
    <td>PixArt-α</td> <td>&#10004</td> <td>68.3</td> <td>60.9</td> <td>93.2</td> <td>77.5</td> <td>45.5</td> 
</tr> 
<tr> 
    <td>Playground 2.5</td> <td>&#10004</td> <td>71.9</td> <td>70.8</td> <td>94.9</td> <td>83.3</td> <td>54.3</td> 
</tr> 

<tr> 
    <td>SD 3</td> <td>&#10008</td> <td>77.1</td> <td>69.3</td> <td>94.6</td> <td>82.5</td> <td>56.7</td> 
    
</tr> 
<tr style="font-weight: bold; background-color: #f2f2f2;"> 
    <td>Hunyuan-DiT</td><td>&#10004</td> <td>74.2</td> <td>74.3</td> <td>95.4</td> <td>86.6</td> <td>59.0</td> 
</tr>
<tr> 
    <td>MidJourney v6</td><td>&#10008</td> <td>73.5</td> <td>80.2</td> <td>93.5</td> <td>87.2</td> <td>63.3</td> 
</tr> 
<tr> 
    <td>DALL-E 3</td><td>&#10008</td> <td>83.9</td> <td>80.3</td> <td>96.5</td> <td>89.4</td> <td>71.0</td> 
</tr> 
</table> 
</p>

## 🎥Visualization

* **Chinese Elements**
<p align="center">
  <img src="./asset/chinese elements understanding.png"  height=220>
</p>

* **Long Text Input**


<p align="center">
  <img src="./asset/long text understanding.png"  height=310>
</p>

* **Multi-turn Text2Image Generation**

https://github.com/yestinl/MDINO/assets/27557933/084ac599-73ce-4be9-9ba9-1b69354f64f8

---

## 📜Requirements

This repo consists of DialogGen (a prompt enhancement model) and Hunyuan-DiT (a text-to-image model).

The following table shows the requirements for running the models (The TensorRT version will be updated soon):

|          Model           | TensorRT | Batch Size | GPU Memory |    GPU    |
|:------------------------:|:--------:|:----------:|:----------:|:---------:|
| DialogGen + Hunyuan-DiT  |    ✘     |     1      |    32G     | V100/A100 |
|       Hunyuan-DiT        |    ✘     |     1      |    11G     | V100/A100 |

<!-- | DialogGen + Hunyuan-DiT  |    ✔     |     1      |     ?      |   A100    |
|       Hunyuan-DiT        |    ✔     |     1      |     ?      |   A100    | -->

* An NVIDIA GPU with CUDA support is required. 
  * We have tested V100 and A100 GPUs.
  * **Minimum**: The minimum GPU memory required is 11GB.
  * **Recommended**: We recommend using a GPU with 32GB of memory for better generation quality.
* Tested operating system: Linux

## 🛠️Dependencies and Installation

Begin by cloning the repository:
```bash
git clone https://github.com/tencent/HunyuanDiT
cd HunyuanDiT
```

We provide an `environment.yml` file for setting up a Conda environment.
Conda's installation instructions are available [here](https://docs.anaconda.com/free/miniconda/index.html).

```bash
# 1. Prepare conda environment
conda env create -f environment.yml

# 2. Activate the environment
conda activate HunyuanDiT

# 3. Install pip dependencies
python -m pip install -r requirements.txt

# 4. (Optional) Install flash attention v2 for acceleration (requires CUDA 11.6 or above)
python -m pip install git+https://github.com/Dao-AILab/flash-attention.git@v2.1.2.post3
```

## 🧱Download Pretrained Models
To download the model, first install the huggingface-cli. (Detailed instructions are available [here](https://huggingface.co/docs/huggingface_hub/guides/cli).)

```bash
python -m pip install "huggingface_hub[cli]"
```

Then download the model using the following commands:

```bash
# Create a directory named 'ckpts' where the model will be saved, fulfilling the prerequisites for running the demo.
mkdir ckpts
# Use the huggingface-cli tool to download the model.
# The download time may vary from 10 minutes to 1 hour depending on network conditions.
huggingface-cli download Tencent-Hunyuan/HunyuanDiT --local-dir ./ckpts
```
Note：If an `No such file or directory: 'ckpts/.huggingface/.gitignore.lock'` like error occurs during the download process, you can ignore the error and retry the command by executing `huggingface-cli download Tencent-Hunyuan/HunyuanDiT --local-dir ./ckpts`

For more information about the model, visit the Hugging Face repository [here](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT).


All models will be automatically downloaded. For more information about the model, visit the Hugging Face repository [here](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT).

|       Model        | #Params |                                              Download URL                                               |
|:------------------:|:-------:|:-------------------------------------------------------------------------------------------------------:|
|        mT5         |  1.6B   |               [mT5](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT/tree/main/t2i/mt5)                |
|        CLIP        |  350M   |        [CLIP](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT/tree/main/t2i/clip_text_encoder)        |
|     DialogGen      |  7.0B   |           [DialogGen](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT/tree/main/dialoggen)            |
| sdxl-vae-fp16-fix  |   83M   | [sdxl-vae-fp16-fix](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT/tree/main/t2i/sdxl-vae-fp16-fix)  |
|    Hunyuan-DiT     |  1.5B   |          [Hunyuan-DiT](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT/tree/main/t2i/model)           |


## 🔑 Inference

### Using Gradio

Make sure you have activated the conda environment before running the following command.

```shell
# By default, we start a Chinese UI.
python app/hydit_app.py

# Using Flash Attention for acceleration.
python app/hydit_app.py --infer-mode fa

# You can disable the enhancement model if the GPU memory is insufficient.
# The enhancement will be unavailable until you restart the app without the `--no-enhance` flag. 
python app/hydit_app.py --no-enhance

# Start with English UI
python app/hydit_app.py --lang en
```

### Using Command Line

We provide 3 modes to quick start: 

```bash
# Prompt Enhancement + Text-to-Image. Torch mode
python sample_t2i.py --prompt "渔舟唱晚"

# Only Text-to-Image. Torch mode
python sample_t2i.py --prompt "渔舟唱晚" --no-enhance

# Only Text-to-Image. Flash Attention mode
python sample_t2i.py --infer-mode fa --prompt "渔舟唱晚"

# Generate an image with other image sizes.
python sample_t2i.py --prompt "渔舟唱晚" --image-size 1280 768
```

### More Configurations

We list some more useful configurations for easy usage:

|    Argument     |  Default  |                     Description                     |
|:---------------:|:---------:|:---------------------------------------------------:|
|   `--prompt`    |   None    |        The text prompt for image generation         |
| `--image-size`  | 1024 1024 |           The size of the generated image           |
|    `--seed`     |    42     |        The random seed for generating images        |
| `--infer-steps` |    100    |          The number of steps for sampling           |
|  `--negative`   |     -     |      The negative prompt for image generation       |
| `--infer-mode`  |   torch   |          The inference mode (torch or fa)           |
|   `--sampler`   |   ddpm    |    The diffusion sampler (ddpm, ddim, or dpmms)     |
| `--no-enhance`  |   False   |        Disable the prompt enhancement model         |
| `--model-root`  |   ckpts   |     The root directory of the model checkpoints     |
|  `--load-key`   |    ema    | Load the student model or EMA model (ema or module) |


# 🔗 BibTeX
If you find Hunyuan-DiT useful for your research and applications, please cite using this BibTeX:

```BibTeX
@misc{hunyuandit,
      title={Hunyuan-DiT: A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding},
      author={Zhimin Li, Jianwei Zhang, Qin Lin, Jiangfeng Xiong, Yanxin Long, Xinchi Deng, Yingfang Zhang, Xingchao Liu, Minbin Huang, Zedong Xiao, Dayou Chen, Jiajun He, Jiahao Li, Wenyue Li, Chen Zhang, Rongwei Quan, Jianxiang Lu, Jiabin Huang, Xiaoyan Yuan, Xiaoxiao Zheng, Yixuan Li, Jihong Zhang, Chao Zhang, Meng Chen, Jie Liu, Zheng Fang, Weiyan Wang, Jinbao Xue, Yangyu Tao, JianChen Zhu, Kai Liu, Sihuan Lin, Yifu Sun, Yun Li, Dongdong Wang, Zhichao Hu, Xiao Xiao, Yan Chen, Yuhong Liu, Wei Liu, Di Wang, Yong Yang, Jie Jiang, Qinglin Lu},
      year={2024},
}
```
Model	Open Source	Text-Image Consistency (%)	Excluding AI Artifacts (%)	Subject Clarity (%)	Aesthetics (%)	Overall (%)
Model	Open Source	SDXL	✔	64.3	Aesthetics (%)	Overall (%)	60.6	91.1	76.3	42.7
PixArt-α	✔	68.3	60.9	93.2	77.5	45.5
Playground 2.5	✔	71.9	70.8	94.9	83.3	54.3
SD 3	✘	77.1	69.3	94.6	82.5	56.7
Hunyuan-DiT	✔	74.2	74.3	95.4	86.6	59.0
MidJourney v6	✘	73.5	80.2	93.5	87.2	63.3
DALL-E 3	✘	83.9	80.3	96.5	89.4	71.0