File size: 1,333 Bytes
8ef5ecf | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 | ---
license: apache-2.0
library_name: transformers
---
# FineTunedBest
<div align="center">
<img src="figures/fig1.png" width="65%" alt="FineTunedBest Architecture" />
</div>
## Overview
FineTunedBest is a fine-tuned RoBERTa model optimized for multi-task performance across reasoning, comprehension, and generation benchmarks. This model was selected from a series of hyperparameter sweeps as the run with the highest evaluation accuracy.
<div align="center">
<img width="75%" src="figures/fig2.png">
</div>
## Benchmark Results
| Benchmark | FineTunedBest |
|---|---|
| Math Reasoning | 0.550 |
| Logical Reasoning | 0.819 |
| Reading Comprehension | 0.700 |
| Code Generation | 0.650 |
| Summarization | 0.767 |
| Instruction Following | 0.758 |
<div align="center">
<img width="70%" src="figures/fig3.png">
</div>
## Training Details
The model was fine-tuned with an optimized hyperparameter configuration discovered through a systematic sweep. See the associated config.json for full details.
## Usage
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer
model = AutoModelForSequenceClassification.from_pretrained("FineTunedBest-TestRepo")
tokenizer = AutoTokenizer.from_pretrained("FineTunedBest-TestRepo")
```
## License
This model is released under the Apache 2.0 license.
|