Update README.md
Browse files
README.md
CHANGED
|
@@ -3,6 +3,37 @@ license: apache-2.0
|
|
| 3 |
library_name: transformers
|
| 4 |
---
|
| 5 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
|
| 7 |
#### Quickstart
|
| 8 |
|
|
|
|
| 3 |
library_name: transformers
|
| 4 |
---
|
| 5 |
|
| 6 |
+
<div align='center'>
|
| 7 |
+
<h1>Emu3: Next-Token Prediction is All You Need</h1h1>
|
| 8 |
+
<h3></h3>
|
| 9 |
+
|
| 10 |
+
[Emu3 Team, BAAI](https://www.baai.ac.cn/english.html)
|
| 11 |
+
|
| 12 |
+
| [Project Page](https://emu.baai.ac.cn) | [Paper](https://baai-solution.ks3-cn-beijing.ksyuncs.com/emu3/Emu3-tech-report.pdf?KSSAccessKeyId=AKLTgew6Kdg6RsK92QSfB2KLA&Expires=2591406552&Signature=6BvwfLVqvfww26Bhwvk3mG0FrL8%3D) | [🤗HF Models](https://huggingface.co/collections/BAAI/emu3-66f4e64f70850ff358a2e60f) | [github](https://github.com/baaivision/Emu3) |
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
</div>
|
| 16 |
+
|
| 17 |
+
<div align='center'>
|
| 18 |
+
<img src="https://github.com/baaivision/Emu3/blob/main/assets/arch.png?raw=True" class="interpolation-image" alt="arch." height="80%" width="70%" />
|
| 19 |
+
</div>
|
| 20 |
+
|
| 21 |
+
We introduce **Emu3**, a new suite of state-of-the-art multimodal models trained solely with **<i>next-token prediction</i>**! By tokenizing images, text, and videos into a discrete space, we train a single transformer from scratch on a mixture of multimodal sequences.
|
| 22 |
+
|
| 23 |
+
### Emu3 excels in both generation and perception
|
| 24 |
+
**Emu3** outperforms several well-established task-specific models in both generation and perception tasks, surpassing flagship open models such as SDXL, LLaVA-1.6 and OpenSora-1.2, while eliminating the need for diffusion or compositional architectures.
|
| 25 |
+
|
| 26 |
+
<div align='center'>
|
| 27 |
+
<img src="https://github.com/baaivision/Emu3/blob/main//assets/comparison.png?raw=True" class="interpolation-image" alt="comparison." height="80%" width="80%" />
|
| 28 |
+
</div>
|
| 29 |
+
|
| 30 |
+
### Highlights
|
| 31 |
+
|
| 32 |
+
- **Emu3** is capable of generating high-quality images following the text input, by simply predicting the next vision token. The model naturally supports flexible resolutions and styles.
|
| 33 |
+
- **Emu3** shows strong vision-language understanding capabilities to see the physical world and provides coherent text responses. Notably, this capability is achieved without depending on a CLIP and a pretrained LLM.
|
| 34 |
+
- **Emu3** simply generates a video causally by predicting the next token in a video sequence, unlike the video diffusion model as in Sora. With a video in context, Emu3 can also naturally extend the video and predict what will happen next.
|
| 35 |
+
|
| 36 |
+
|
| 37 |
|
| 38 |
#### Quickstart
|
| 39 |
|