Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
datasets:
|
| 3 |
+
- Aria-UI/Aria-UI_Data
|
| 4 |
+
base_model:
|
| 5 |
+
- rhymes-ai/Aria-Base-8K
|
| 6 |
+
- rhymes-ai/Aria
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
<div style="display: flex; align-items: center; gap: 20px;">
|
| 10 |
+
<img src="assets/performance_spider.pdf" alt="Performance Spider Plot" width="400">
|
| 11 |
+
<!-- Text Section -->
|
| 12 |
+
<div>
|
| 13 |
+
<h2>Key Features of Aria-UI</h2>
|
| 14 |
+
<div style="font-weight: bold;">✨ Versatile Grounding Instruction Understanding:</div>
|
| 15 |
+
<div>Aria-UI handles diverse grounding instructions, excelling in interpreting varied formats, ensuring robust adaptability across dynamic scenarios or when paired with diverse planning agents.</div>
|
| 16 |
+
<div style="font-weight: bold; margin-top: 10px;">📝 Context-aware Grounding:</div>
|
| 17 |
+
<div>Aria-UI effectively leverages historical input, whether in pure text or text-image-interleaved formats, to improve grounding accuracy.</div>
|
| 18 |
+
<div style="font-weight: bold; margin-top: 10px;">⚡ Lightweight and Fast:</div>
|
| 19 |
+
<div>Aria-UI is a mixture-of-expert model with 3.9B activated parameters per token. It efficiently encodes GUI input of variable sizes and aspect ratios, with ultra-resolution support.</div>
|
| 20 |
+
<div style="font-weight: bold; margin-top: 10px;">🎉 Superior Performances:</div>
|
| 21 |
+
<div>
|
| 22 |
+
Aria-UI sets new state-of-the-art results on offline and online agent benchmarks.
|
| 23 |
+
<div style="margin-top: 5px;">
|
| 24 |
+
<span style="font-weight: bold;">🏆 1st place</span> on <span style="font-weight: bold;">AndroidWorld</span> with <span style="font-weight: bold;">44.8%</span> task success rate and
|
| 25 |
+
<span style="font-weight: bold;">🥉 3rd place</span> on <span style="font-weight: bold;">OSWorld</span> with <span style="font-weight: bold;">15.2%</span> task success rate (Dec. 2024).
|
| 26 |
+
</div>
|
| 27 |
+
</div>
|
| 28 |
+
</div>
|
| 29 |
+
</div>
|