File size: 3,042 Bytes
f751e98
 
 
2c77c00
add998a
2c77c00
 
 
 
307f67d
 
 
2c77c00
 
 
 
 
 
 
c32c4c1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03da1c5
c32c4c1
03da1c5
c32c4c1
 
03da1c5
c32c4c1
03da1c5
c32c4c1
 
2c77c00
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
307f67d
 
 
 
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
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
---
license: apache-2.0
---

# Model card for climateBUG-LM

## Model Description
climateBUG-LM is a deep learning language model fine-tuned for analyzing bank reports in the context of climate change and sustainability. It leverages a unique annotated corpus, climateBUG-Data, which consists of statements from EU banks' annual and sustainability reports, focusing on climate change and finance. This model aims to classify statements as relevant or irrelevant to climate-related subjects, offering enhanced performance due to its domain-specific training.

## Access and Usage
- Models, dataset and tools are available at the [climateBUG project page](https://www.climatebug.se/).
- Suitable for researchers and professionals in finance, sustainability, and climate policy.

## Applications
The model is ideal for:
+ Analyzing financial reports for climate change-related content.
+ Research in financial sustainability and climate economics.
+ Tracking how banks articulate their climate-related activities.

## Example Usage
Here is an example of how to use the climateBUG-LM model for classifying text as climate-related or not:

```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline

tokenizer_name = "lumilogic/climateBUG-LM"
model_name = "lumilogic/climateBUG-LM"


model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(tokenizer_name, max_len=512)

pipe = pipeline("text-classification", model=model,
                tokenizer=tokenizer, device_map='auto')


# Climate related text
text = 'This issue represents around 10% of the outstanding volume of green sovereign bonds and will used to finance Germany’s climate and environmental strategy.'
output = pipe(text)
print(output)  # [{'label': 'LABEL_1', 'score': 0.9974282383918762}]

# Non-climate related text
text = 'Our model, based on a customer-centric universal banking relationship, therefore demonstrated its resilience and usefulness for all stakeholders in all our regions.'
output = pipe(text)
print(output)  # [{'label': 'LABEL_0', 'score': 0.9931207299232483}]
```

## Limitations
+ Optimized for EU bank reports; performance may vary for other regions.
+ Primarily focused on climate and finance domains.

## Citation
Please cite this model as follows:

Yu, Y., Scheidegger, S., Elliott, J., & Löfgren, Å. (2024). climateBUG: A data-driven framework for analyzing bank reporting through a climate lens. Expert Systems With Applications, 239, 122162.

```bibtex
@article{yu2024climatebug,
title = {climateBUG : A data-driven framework for analyzing bank reporting through a climate lens},
journal = {Expert Systems with Applications},
volume = {239},
pages = {122162},
year = {2024},
author = {Yinan Yu and Samuel Scheidegger and Jasmine Elliott and Åsa Löfgren}
}
```

## Support and Contact
For support, additional information, or inquiries, please reach out through [email protected] or visit the [climateBUG project page](https://www.climatebug.se/).