---
library_name: transformers
license: apache-2.0
base_model: intfloat/multilingual-e5-base
tags:
- generated_from_trainer
- sentence-transformers
- text-classification
- feature-extraction
- generated_from_trainer
- legal
- taxation
- fiscalité
- tax
metrics:
- accuracy
model-index:
- name: lemone-router
  results: []
language:
- fr
pipeline_tag: text-classification
datasets:
- louisbrulenaudet/code-impots
- louisbrulenaudet/code-impots-annexe-iv
- louisbrulenaudet/code-impots-annexe-iii
- louisbrulenaudet/code-impots-annexe-i
- louisbrulenaudet/code-impots-annexe-ii
- louisbrulenaudet/livre-procedures-fiscales
- louisbrulenaudet/bofip
---

<img src="assets/thumbnail.webp">

# Lemone-Router: A Series of Fine-Tuned Classification Models for French Taxation

This model is a fine-tuned version of [intfloat/multilingual-e5-base](https://huggingface.co/intfloat/multilingual-e5-base).
It achieves the following results on the evaluation set:
- Loss: 0.4096
- Accuracy: 0.9265

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 4.099463734610582e-05
- train_batch_size: 16
- eval_batch_size: 64
- seed: 23
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.5371        | 1.0   | 2809  | 0.4147          | 0.8680   |
| 0.3154        | 2.0   | 5618  | 0.3470          | 0.8914   |
| 0.2241        | 3.0   | 8427  | 0.3345          | 0.9147   |
| 0.1273        | 4.0   | 11236 | 0.3788          | 0.9187   |
| 0.0525        | 5.0   | 14045 | 0.4096          | 0.9265   |

### Training Hardware
- **On Cloud**: No
- **GPU Model**: 1 x NVIDIA H100 NVL
- **CPU Model**: AMD EPYC 9V84 96-Core Processor
- **RAM Size**: 314.68 GB

### Framework versions

- Transformers 4.45.2
- Pytorch 2.4.1+cu121
- Datasets 2.21.0
- Tokenizers 0.20.1

## Citation
If you use this code in your research, please use the following BibTeX entry.

```BibTeX
@misc{louisbrulenaudet2024,
  author =       {Louis Brulé Naudet},
  title =        {Lemone-Embed: A Series of Fine-Tuned Embedding Models for French Taxation},
  year =         {2024}
  howpublished = {\url{https://huggingface.co/datasets/louisbrulenaudet/lemone-embed-pro}},
}
```

## Feedback

If you have any feedback, please reach out at [louisbrulenaudet@icloud.com](mailto:louisbrulenaudet@icloud.com).