Text Classification
Transformers
PyTorch
bert
Inference Endpoints
File size: 3,447 Bytes
53f7300
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
---

language:
  - de
  - en
  - es
  - fr
---


# Model Card for `passage-ranker-v1-XS-multilingual`

This model is a passage ranker developed by Sinequa. It produces a relevance score given a query-passage pair and is
used to order search results.

Model name: `passage-ranker-v1-XS-multilingual`

## Supported Languages

The model was trained and tested in the following languages:

- English
- French
- German
- Spanish

## Scores

| Metric              | Value |
|:--------------------|------:|
| Relevance (NDCG@10) | 0.456 |

Note that the relevance score is computed as an average over 14 retrieval datasets (see
[details below](#evaluation-metrics)).

## Inference Times

| GPU        | Batch size 32 |
|:-----------|--------------:|
| NVIDIA A10 |          4 ms |
| NVIDIA T4  |         13 ms |

The inference times only measure the time the model takes to process a single batch, it does not include pre- or
post-processing steps like the tokenization. The reported times are measured using the
[FP16](https://en.wikipedia.org/wiki/Half-precision_floating-point_format) version of the model.

## Requirements

- Minimal Sinequa version: 11.10.0
- GPU memory usage: 300 MiB

Note that GPU memory usage only includes how much GPU memory the actual model consumes on an NVIDIA T4 GPU with a batch
size of 32. It does not include the fix amount of memory that is consumed by the ONNX Runtime upon initialization which
can be around 0.5 to 1 GiB depending on the used GPU.

## Model Details

### Overview

- Number of parameters: 16 million
- Base language model: Homegrown Sinequa BERT-Mini ([Paper](https://arxiv.org/abs/1908.08962)) pretrained in the four
  supported languages
- Insensitive to casing and accents
- Training procedure: [MonoBERT](https://arxiv.org/abs/1901.04085)

### Training Data

- MS MARCO Passage Ranking
  ([Paper](https://arxiv.org/abs/1611.09268),
  [Official Page](https://microsoft.github.io/msmarco/),
  [English & translated datasets on the HF dataset hub](https://huggingface.co/datasets/unicamp-dl/mmarco))
    - Original English dataset
    - Translated datasets for the other three supported languages

### Evaluation Metrics

To determine the relevance score, we averaged the results that we obtained when evaluating on the datasets of the
[BEIR benchmark](https://github.com/beir-cellar/beir). Note that all these datasets are in English.

| Dataset           | NDCG@10 |
|:------------------|--------:|
| Average           |   0.456 |
|                   |         |
| Arguana           |   0.517 |
| CLIMATE-FEVER     |   0.159 |
| DBPedia Entity    |   0.355 |
| FEVER             |   0.733 |
| FiQA-2018         |   0.282 |
| HotpotQA          |   0.688 |
| MS MARCO          |   0.327 |
| NFCorpus          |   0.341 |
| NQ                |   0.441 |
| Quora             |   0.768 |
| SCIDOCS           |   0.143 |
| SciFact           |   0.629 |
| TREC-COVID        |   0.667 |
| Webis-Touche-2020 |   0.328 |

We evaluated the model on the datasets of the [MIRACL benchmark](https://github.com/project-miracl/miracl) to test its
multilingual capacities. Note that not all training languages are part of the benchmark, so we only report the metrics
for the existing languages.

| Language | NDCG@10 |
|:---------|--------:|
| French   |   0.349 |
| German   |   0.375 |
| Spanish  |   0.417 |