Update README.md
Browse files
README.md
CHANGED
|
@@ -78,6 +78,20 @@ Safety alignment is another aspect we particularly emphasize. The Web3 investmen
|
|
| 78 |
|
| 79 |
## Performance Metrics
|
| 80 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 81 |
|
| 82 |
|
| 83 |
## Application Scenarios
|
|
@@ -127,10 +141,10 @@ This model follows the Apache 2.0 open-source license. Commercial use must compl
|
|
| 127 |
## Citation
|
| 128 |
|
| 129 |
```bibtex
|
| 130 |
-
@misc{
|
| 131 |
title={DMind-2: Advanced Web3 Domain-Specific Large Language Models with Distribution-Preserving CoT Distillation},
|
| 132 |
author={DMind Team},
|
| 133 |
-
year={
|
| 134 |
publisher={Hugging Face}
|
| 135 |
}
|
| 136 |
```
|
|
@@ -138,7 +152,6 @@ This model follows the Apache 2.0 open-source license. Commercial use must compl
|
|
| 138 |
## Contact
|
| 139 |
|
| 140 |
- 🌐 Project Homepage: [https://dmind.ai](https://dmind.ai)
|
| 141 |
-
- 📧 Technical Support: tech@dmind.ai
|
| 142 |
- 💬 Community Discussion: [Discord](https://discord.gg/dmind)
|
| 143 |
- 🐦 Twitter: [@DMindAI](https://twitter.com/DMindAI)
|
| 144 |
|
|
|
|
| 78 |
|
| 79 |
## Performance Metrics
|
| 80 |
|
| 81 |
+
| Category | Benchmark (Metric) | DeepSeek-R1-0528-Qwen3-8B | gpt-oss-20b | Qwen3-32B | Qwen3-4B(Thinking) | **DMind2-mini(4B)** |
|
| 82 |
+
| :--- | :--- | :--- | :--- | :--- | :--- | :--- |
|
| 83 |
+
| **General** | | | | | | |
|
| 84 |
+
| | MMLU-Pro (EM) | 84.0 | 85.3 | 65.54 | 70.4 | 82.1 |
|
| 85 |
+
| | GPQA-Diamond (Pass@1) | 71.5 | 71.5 | 68.4 | 55.9 | 64.2 |
|
| 86 |
+
| | SimpleQA (Correct) | 30.1 | 6.7 | 8.0 | 1.0 | - |
|
| 87 |
+
| **Math** | | | | | | |
|
| 88 |
+
| | AIME 2024 (Pass@1) | 79.8 | 96 | 81.4 | - | 92.7 |
|
| 89 |
+
| | AIME 2025 (Pass@1) | 70.0 | 98.7 | 72.9 | 65.6 | 81.6 |
|
| 90 |
+
| | CNMO 2024 (Pass@1) | 78.8 | 86.9 | - | - | 82.4 |
|
| 91 |
+
| **Tools** | | | | | | |
|
| 92 |
+
| | BFCL_v3 | - | 67.8 | 70.3 | 65.9 | 70.2 |
|
| 93 |
+
| **Web3** | | | | | | |
|
| 94 |
+
| | DMind Benchmark | 71.5 | 70.2 | 68.0 | 74.5 | 80.1 |
|
| 95 |
|
| 96 |
|
| 97 |
## Application Scenarios
|
|
|
|
| 141 |
## Citation
|
| 142 |
|
| 143 |
```bibtex
|
| 144 |
+
@misc{dmind2025,
|
| 145 |
title={DMind-2: Advanced Web3 Domain-Specific Large Language Models with Distribution-Preserving CoT Distillation},
|
| 146 |
author={DMind Team},
|
| 147 |
+
year={2025},
|
| 148 |
publisher={Hugging Face}
|
| 149 |
}
|
| 150 |
```
|
|
|
|
| 152 |
## Contact
|
| 153 |
|
| 154 |
- 🌐 Project Homepage: [https://dmind.ai](https://dmind.ai)
|
|
|
|
| 155 |
- 💬 Community Discussion: [Discord](https://discord.gg/dmind)
|
| 156 |
- 🐦 Twitter: [@DMindAI](https://twitter.com/DMindAI)
|
| 157 |
|