Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
library_name: transformers
|
| 4 |
+
---
|
| 5 |
+
# Titan-Math-Pro
|
| 6 |
+
<!-- markdownlint-disable first-line-h1 -->
|
| 7 |
+
<!-- markdownlint-disable html -->
|
| 8 |
+
<!-- markdownlint-disable no-duplicate-header -->
|
| 9 |
+
|
| 10 |
+
<div align="center">
|
| 11 |
+
<img src="figures/fig1.png" width="60%" alt="Titan-Math-Pro" />
|
| 12 |
+
</div>
|
| 13 |
+
<hr>
|
| 14 |
+
|
| 15 |
+
<div align="center" style="line-height: 1;">
|
| 16 |
+
<a href="LICENSE" style="margin: 2px;">
|
| 17 |
+
<img alt="License" src="figures/fig2.png" style="display: inline-block; vertical-align: middle;"/>
|
| 18 |
+
</a>
|
| 19 |
+
</div>
|
| 20 |
+
|
| 21 |
+
## 1. Introduction
|
| 22 |
+
|
| 23 |
+
Titan-Math-Pro is a mathematics-first model trained on verified proofs and competition problems with process supervision.
|
| 24 |
+
|
| 25 |
+
## 2. Evaluation Results
|
| 26 |
+
|
| 27 |
+
### Comprehensive Benchmark Results
|
| 28 |
+
|
| 29 |
+
<div align="center">
|
| 30 |
+
|
| 31 |
+
| | Benchmark | NumeriX | Titan-base | ProofNet | Titan-Math-Pro |
|
| 32 |
+
|---|---|---|---|---|---|
|
| 33 |
+
| **Core Reasoning Tasks** | Math Reasoning | 0.589 | 0.595 | 0.577 | 0.633 |
|
| 34 |
+
| | Logical Reasoning | 0.816 | 0.818 | 0.793 | 0.850 |
|
| 35 |
+
| | Common Sense | 0.749 | 0.735 | 0.776 | 0.783 |
|
| 36 |
+
| **Language Understanding** | Reading Comprehension | 0.707 | 0.731 | 0.743 | 0.760 |
|
| 37 |
+
| | Question Answering | 0.595 | 0.595 | 0.642 | 0.649 |
|
| 38 |
+
| | Text Classification | 0.815 | 0.817 | 0.831 | 0.849 |
|
| 39 |
+
| | Sentiment Analysis | 0.779 | 0.780 | 0.765 | 0.818 |
|
| 40 |
+
| **Generation Tasks** | Code Generation | 0.703 | 0.682 | 0.702 | 0.730 |
|
| 41 |
+
| | Creative Writing | 0.693 | 0.670 | 0.658 | 0.700 |
|
| 42 |
+
| | Dialogue Generation | 0.683 | 0.677 | 0.656 | 0.702 |
|
| 43 |
+
| | Summarization | 0.793 | 0.765 | 0.771 | 0.805 |
|
| 44 |
+
| **Specialized Capabilities** | Translation | 0.802 | 0.815 | 0.814 | 0.826 |
|
| 45 |
+
| | Knowledge Retrieval | 0.667 | 0.672 | 0.687 | 0.718 |
|
| 46 |
+
| | Instruction Following | 0.756 | 0.750 | 0.741 | 0.798 |
|
| 47 |
+
| | Safety Evaluation | 0.762 | 0.739 | 0.734 | 0.779 |
|
| 48 |
+
|
| 49 |
+
</div>
|
| 50 |
+
|
| 51 |
+
### Overall Performance Summary
|
| 52 |
+
The Titan-Math-Pro demonstrates strong performance across all evaluated benchmark categories, with particularly notable results in reasoning and generation tasks.
|
| 53 |
+
|
| 54 |
+
## 3. Chat Website & API Platform
|
| 55 |
+
We offer a chat interface and API for you to interact with Titan-Math-Pro. Please check our official website for more details.
|
| 56 |
+
|
| 57 |
+
## 4. How to Run Locally
|
| 58 |
+
|
| 59 |
+
Please refer to our code repository for more information about running Titan-Math-Pro locally.
|
| 60 |
+
|
| 61 |
+
### Temperature
|
| 62 |
+
We recommend setting the temperature parameter to 0.6.
|
| 63 |
+
|
| 64 |
+
## 5. License
|
| 65 |
+
This repository is released under the mit license. The model supports commercial use.
|
| 66 |
+
|
| 67 |
+
## 6. Contact
|
| 68 |
+
If you have any questions, please contact us at math@titan-compute.com.
|