File size: 3,231 Bytes
75025cc
69c4876
 
 
 
 
75025cc
69c4876
75025cc
69c4876
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d0570f4
 
 
 
 
 
 
 
 
 
 
 
 
 
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
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
---
pipeline_tag: text-classification
tags:
- transformers
- information-retrieval
language: pl
license: apache-2.0

---

<h1 align="center">polish-reranker-large-mse</h1>

This is a Polish text ranking model trained using the mean squared error (MSE) distillation method on a large dataset of text pairs consisting of 1.4 million queries and 10 million documents. 
The training data included the following parts: 1) The Polish MS MARCO training split (800k queries); 2) The ELI5 dataset translated to Polish (over 500k queries); 3) A collection of Polish medical questions and answers (approximately 100k queries).
As a teacher model, we employed [unicamp-dl/mt5-13b-mmarco-100k](https://huggingface.co/unicamp-dl/mt5-13b-mmarco-100k), a large multilingual reranker based on the MT5-XXL architecture. As a student model, we choose [Polish RoBERTa](https://huggingface.co/sdadas/polish-roberta-large-v2).
In the MSE method, the student is trained to directly replicate the outputs returned by the teacher.

## Usage (Sentence-Transformers)

You can use the model like this with [sentence-transformers](https://www.SBERT.net):

```python
from sentence_transformers import CrossEncoder
import torch.nn

query = "Jak dożyć 100 lat?"
answers = [
    "Trzeba zdrowo się odżywiać i uprawiać sport.",
    "Trzeba pić alkohol, imprezować i jeździć szybkimi autami.",
    "Gdy trwała kampania politycy zapewniali, że rozprawią się z zakazem niedzielnego handlu."
]

model = CrossEncoder(
    "sdadas/polish-reranker-large-mse",
    default_activation_function=torch.nn.Identity(),
    max_length=512,
    device="cuda" if torch.cuda.is_available() else "cpu"
)
pairs = [[query, answer] for answer in answers]
results = model.predict(pairs)
print(results.tolist())
```

## Usage (Huggingface Transformers)

The model can also be used with Huggingface Transformers in the following way:

```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import numpy as np

query = "Jak dożyć 100 lat?"
answers = [
    "Trzeba zdrowo się odżywiać i uprawiać sport.",
    "Trzeba pić alkohol, imprezować i jeździć szybkimi autami.",
    "Gdy trwała kampania politycy zapewniali, że rozprawią się z zakazem niedzielnego handlu."
]

model_name = "sdadas/polish-reranker-large-mse"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)
texts = [f"{query}</s></s>{answer}" for answer in answers]
tokens = tokenizer(texts, padding="longest", max_length=512, truncation=True, return_tensors="pt")
output = model(**tokens)
results = output.logits.detach().numpy()
results = np.squeeze(results)
print(results.tolist())
```

## Evaluation Results

The model achieves **NDCG@10** of **60.27** in the Rerankers category of the Polish Information Retrieval Benchmark. See [PIRB Leaderboard](https://huggingface.co/spaces/sdadas/pirb) for detailed results.

## Citation

```bibtex
@article{dadas2024assessing,
  title={Assessing generalization capability of text ranking models in Polish}, 
  author={Sławomir Dadas and Małgorzata Grębowiec},
  year={2024},
  eprint={2402.14318},
  archivePrefix={arXiv},
  primaryClass={cs.CL}
}
```