BramVanroy/fietje-2

Fietje 2

An open and efficient LLM for Dutch

👱‍♀️ Base version (this one) - 🤖 Instruct version - 💬 Chat version - 🚀 GGUF of base

Chat with Fietje here!

Fietje is an adapated version of microsoft/phi-2, tailored to Dutch text generation by training on 28B tokens. It is small and efficient with a size of 2.7 billion parameters while performing almost on par with more powerful Dutch LLMs of twice its size like GEITje 7B Ultra.

A thorough description of the creation and evaluation of Fietje as well as usage examples are available in this Github repository.

Citation

If you use Fietje or the CulturaX + Wikipedia filtered subset in your work, please cite to the following paper:

@misc{vanroy2024fietjeopenefficientllm,
      title={Fietje: An open, efficient LLM for Dutch}, 
      author={Bram Vanroy},
      year={2024},
      eprint={2412.15450},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2412.15450}, 
}

Intended uses & limitations

The same limitations as phi-2, and LLMs in general, apply here. LLMs hallucinate, make mistakes, and should not be trusted. Use at your own risk!

Training data

Fietje was continue-pretrained on 28B Dutch tokens, which includes the full Dutch component of Wikipedia (accounting for around 15%), supplemented with Dutch tokens from CulturaX. A newer version of this dataset can be found here, which also describes the filtering that took place to ensure high data quality.

Training procedure

I am thankful to the Flemish Supercomputer Center (VSC) for providing the computational power to accomplish this project. Accounting for waiting for jobs, training took around two weeks on four nodes of 4x A100 80GB each (16 total).

Training was done with the wonderful alignment-handbook, using DeepSpeed as a back-end. Exact training recipes and SLURM script are given in the Github repository.

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 9e-05
• train_batch_size: 40
• eval_batch_size: 40
• seed: 42
• distributed_type: multi-GPU
• num_devices: 16
• gradient_accumulation_steps: 3
• total_train_batch_size: 1920
• total_eval_batch_size: 640
• optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-07
• lr_scheduler_type: linear
• num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
1.6334	0.13	900	1.5937
1.5469	0.26	1800	1.5051
1.4937	0.4	2700	1.4628
1.4633	0.53	3600	1.4375
1.4485	0.66	4500	1.4203
1.4374	0.79	5400	1.4085
1.4278	0.92	6300	1.4013

Framework versions

• Transformers 4.39.1
• Pytorch 2.1.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2