File size: 1,323 Bytes
c959e67 |
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 |
---
license: cc-by-nc-4.0
language:
- en
---
## Introduction
The model is used to evaluate the quality of a candidate patent claim compared to the gold claim.
## Example Usage
```python
from transformers import AutoModel, AutoTokenizer
import torch
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
tokenizer = AutoTokenizer.from_pretrained("lj408/PatClaimEval-Quality", trust_remote_code=True)
model = AutoModel.from_pretrained("lj408/PatClaimEval-Quality", trust_remote_code=True).to(device)
gold_claim = "1. A computer-implemented method comprising: identifying a primary code segment; ..."
candidate_claim = "1. A computer-implemented method for managing logger source code segments in a source code development platform, ..."
res = model.score_pair(gold_claim, candidate_claim , tokenizer, device)
print(res)
```
Note that this evaluation method can serve as a reference, but it may not be accurate in all cases. Users should rely on evaluations by patent professionals for more precise results.
## Citation
If you use this model or code, please cite our paper:
```
@article{jiang2025towards,
title={Towards Better Evaluation for Generated Patent Claims},
author={Jiang, Lekang and Scherz, Pascal A and Goetz, Stephan},
journal={arXiv preprint arXiv:2505.11095},
year={2025}
}
``` |