Update README.md
Browse files
README.md
CHANGED
|
@@ -21,6 +21,16 @@ For more details, refer to our **paper on ArXiv**: [🔗 PromptCoT: Synthesizing
|
|
| 21 |
|
| 22 |
---
|
| 23 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
|
| 25 |
## 🔥 **Quick Start: Using the Model**
|
| 26 |
|
|
@@ -85,7 +95,7 @@ print(outputs[0].outputs[0].text)
|
|
| 85 |
---
|
| 86 |
|
| 87 |
## 🔗 **Full Usage & Advanced Options**
|
| 88 |
-
For advanced usage, including
|
| 89 |
🔹 [GitHub: PromptCoT](https://github.com/zhaoxlpku/PromptCoT)
|
| 90 |
|
| 91 |
---
|
|
|
|
| 21 |
|
| 22 |
---
|
| 23 |
|
| 24 |
+
## 🏆 State-of-the-Art Performance
|
| 25 |
+
**PromptCoT-QwQ-32B** has achieved remarkable results, outperforming all competitors across key benchmarks focused on mathematical reasoning:
|
| 26 |
+
|
| 27 |
+
| **Model** | **GSM8K** | **MATH-500** | **AIME2024** | **AIME2025** |
|
| 28 |
+
| --- | --- | --- | --- | --- |
|
| 29 |
+
| **S1-32B** | - | 93.0% | 56.7% | 26.6% |
|
| 30 |
+
| **LIMO-32B** | - | 94.8% | 57.1% | 46.6% |
|
| 31 |
+
| **QwQ-32B** | - | - | 82.1% | 70.8% |
|
| 32 |
+
| **PromptCoT-QwQ-32B** (**ours**) | 🔥 **96.4% ± 0.2%** | 🔥 **96.7% ± 0.5%** | 🔥 **83.8% ± 2.8%** | 🔥 **75.4% ± 4.7%** |
|
| 33 |
+
|
| 34 |
|
| 35 |
## 🔥 **Quick Start: Using the Model**
|
| 36 |
|
|
|
|
| 95 |
---
|
| 96 |
|
| 97 |
## 🔗 **Full Usage & Advanced Options**
|
| 98 |
+
For advanced usage, including batch inference and evaluation on mathematical benchmarks, refer to the **full repository on GitHub**:
|
| 99 |
🔹 [GitHub: PromptCoT](https://github.com/zhaoxlpku/PromptCoT)
|
| 100 |
|
| 101 |
---
|