---
language:
  - fr
license: apache-2.0
tags:
  - medical
  - healthcare
  - french
  - checkpoints
  - composer
  - doctobert
---

# DoctoBERT-fr — Training Checkpoints

<center>
  <img src="assets/logo_doctobert.png" width="100%" alt="DoctoBERT">
</center>

<p align="center">
<a href="https://huggingface.co/blog/bofenghuang/doctobert-fr-release">🤗 Blog</a> |
<a href="https://arxiv.org/abs/2606.22079">📄 Paper</a> |
<a href="https://github.com/doctolib-lab/doctobert">💻 Code</a> |
<a href="https://huggingface.co/collections/doctolib-lab/finemed-fr">🌐 FineMed</a> |
<a href="https://huggingface.co/collections/doctolib-lab/doctobert-fr">🩺 DoctoBERT</a>
</p>

## 📚 Introduction

Per-stage **[MosaicML Composer](https://github.com/mosaicml/composer)** training-state checkpoints for the DoctoBERT-fr encoder family.

Each `.pt` is the full state (model, optimizer, scheduler, RNG), for resuming or analysis. DoctoBERT-fr keeps the final checkpoint of each stage; DoctoModernBERT-fr also keeps intermediate checkpoints every ~5,000 batches.

For ready-to-use inference models, use [DoctoBERT-fr-base](https://huggingface.co/doctolib-lab/doctobert-fr-base) and [DoctoModernBERT-fr-base](https://huggingface.co/doctolib-lab/doctomodernbert-fr-base).

<!--
## 📖 Citation

```bibtex
@misc{doctobert2026,
  title         = {Where Does the Signal Live? A Web Data Recipe for Medical Encoder Pretraining},
  author        = {Huang, Bofeng and Sun, Jacques and Bouchacourt, Diane and Barascud, Nicolas and Fogel, Fajwel},
  year          = {2026},
  eprint        = {2606.XXXXX},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL}
}
```
-->

## ⚖️ License

Released under Apache-2.0. Trained on FineMed-fr and FineMed-rephrased-fr, which derive from FineWeb-2 / FinePDFs (ODC-BY 1.0) and FineWiki (CC BY-SA 4.0); please attribute those upstream sources.

## 🏛️ Acknowledgments

This work was granted access to the HPC resources of IDRIS (Jean Zay) under the allocations 2025-AD011016291 and 2026-A0200617487 made by GENCI.