README.md

---
language: de
license: apache-2.0
datasets: oscar-corpus/OSCAR-2301
---

# mistral7b-de-pure-bf16

Mistral-7B-v0.1 adapted to German as part of our study on efficient language adaptation: "Language Adaptation on a Tight Academic Compute Budget: Tokenizer Swapping Works and Pure bfloat16 Is Enough".

Code: https://github.com/konstantinjdobler/tight-budget-llm-adaptation

Paper: https://openreview.net/forum?id=VYfJaHeVod

## Usage
```python
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("konstantindobler/mistral7b-de-pure-bf16")
model = AutoModelForCausalLM.from_pretrained("konstantindobler/mistral7b-de-pure-bf16")

# Use model and tokenizer as usual
```

## Details
The model is based on [Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) and was adapted to German.
The original tokenizer was kept. 
The model was then trained on 8 billion German tokens from [oscar-corpus/OSCAR-2301](https://huggingface.co/oscar-corpus/OSCAR-2301) with pure bfloat16 precision (no mixed precision). More details and hyperparameters can be found [in the paper](https://openreview.net/forum?id=VYfJaHeVod).

## Disclaimer
The web-scale dataset used for pretraining and tokenizer training ([oscar-corpus/OSCAR-2301](https://huggingface.co/oscar-corpus/OSCAR-2301)) might contain personal and sensitive information.
Such behavior needs to be assessed carefully before any real-world deployment of the models.

## Citation
Please cite as follows:

```bibtex
@inproceedings{dobler2024language,
    title={Language Adaptation on a Tight Academic Compute Budget: Tokenizer Swapping Works and Pure bfloat16 Is Enough},
    author={Konstantin Dobler and Gerard de Melo},
    booktitle={2nd Workshop on Advancing Neural Network Training: Computational Efficiency, Scalability, and Resource Optimization (WANT@ICML 2024)},
    year={2024},
    url={https://openreview.net/forum?id=VYfJaHeVod}
}
```