Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_type: bert
|
| 5 |
+
architectures:
|
| 6 |
+
- BertModel
|
| 7 |
+
tags:
|
| 8 |
+
- bert
|
| 9 |
+
- reasoning
|
| 10 |
+
- code-generation
|
| 11 |
+
- language-model
|
| 12 |
+
- mit
|
| 13 |
+
- benchmarked
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# MyAwesomeModel
|
| 17 |
+
|
| 18 |
+
## Model Information
|
| 19 |
+
**Best Checkpoint**: step_1000 (highest eval_accuracy: 0.875)
|
| 20 |
+
**Overall Weighted Score**: 0.800
|
| 21 |
+
|
| 22 |
+
The MyAwesomeModel has undergone a significant version upgrade. In the latest update, MyAwesomeModel has significantly improved its depth of reasoning and inference capabilities by leveraging increased computational resources and introducing algorithmic optimization mechanisms during post-training.
|
| 23 |
+
|
| 24 |
+
## Key Improvements
|
| 25 |
+
- Significantly improved reasoning capabilities (AIME 2025 accuracy increased from 70% to 87.5%)
|
| 26 |
+
- Reduced hallucination rate
|
| 27 |
+
- Enhanced support for function calling
|
| 28 |
+
- Supports system prompts
|
| 29 |
+
- No special tokens required at output beginning
|
| 30 |
+
|
| 31 |
+
## Comprehensive Benchmark Evaluation Results (All 15 Benchmarks - 3 Decimal Places)
|
| 32 |
+
|
| 33 |
+
| Category | Benchmark | Score (3 decimals) |
|
| 34 |
+
|----------|-----------|-------------------|
|
| 35 |
+
| **Core Reasoning Tasks** | Math Reasoning | 0.875 |
|
| 36 |
+
| | Logical Reasoning | 0.842 |
|
| 37 |
+
| | Common Sense | 0.789 |
|
| 38 |
+
| **Language Understanding** | Reading Comprehension | 0.756 |
|
| 39 |
+
| | Question Answering | 0.723 |
|
| 40 |
+
| | Text Classification | 0.867 |
|
| 41 |
+
| | Sentiment Analysis | 0.834 |
|
| 42 |
+
| **Generation Tasks** | Code Generation | 0.781 |
|
| 43 |
+
| | Creative Writing | 0.712 |
|
| 44 |
+
| | Dialogue Generation | 0.768 |
|
| 45 |
+
| | Summarization | 0.825 |
|
| 46 |
+
| **Specialized Capabilities**| Translation | 0.847 |
|
| 47 |
+
| | Knowledge Retrieval | 0.753 |
|
| 48 |
+
| | Instruction Following | 0.819 |
|
| 49 |
+
| | Safety Evaluation | 0.794 |
|
| 50 |
+
|
| 51 |
+
## Performance Summary
|
| 52 |
+
The MyAwesomeModel demonstrates strong performance across all evaluated benchmark categories, with particularly notable results in reasoning and generation tasks.
|
| 53 |
+
|
| 54 |
+
## Usage
|
| 55 |
+
System prompt recommendation:
|
| 56 |
+
```
|
| 57 |
+
You are MyAwesomeModel, a helpful AI assistant.
|
| 58 |
+
Today is {current date}.
|
| 59 |
+
```
|
| 60 |
+
|
| 61 |
+
Recommended temperature: 0.6
|
| 62 |
+
|
| 63 |
+
## License
|
| 64 |
+
MIT License - supports commercial use and distillation.
|