sqi_latn_1000mb

Goldfish is a suite of monolingual language models trained for 350 languages. This model is the Albanian (Latin script) model trained on 1000MB of data, after accounting for an estimated byte premium of 1.34; content-matched text in Albanian takes on average 1.34x as many UTF-8 bytes to encode as English. The Goldfish models are trained primarily for comparability across languages and for low-resource languages; Goldfish performance for high-resource languages is not designed to be comparable with modern large language models (LLMs).

Note: sqi_latn is a macrolanguage code. Individual language code als_latn (Tosk Albanian) is included in Goldfish, although with less data.

All training and hyperparameter details are in our paper, Goldfish: Monolingual Language Models for 350 Languages (Chang et al., 2024).

Training code and sample usage: https://github.com/tylerachang/goldfish

Sample usage also in this Google Colab: link

Model details:

To access all Goldfish model details programmatically, see https://github.com/tylerachang/goldfish/blob/main/model_details.json. All models are trained with a [CLS] (same as [BOS]) token prepended, and a [SEP] (same as [EOS]) token separating sequences. For best results, make sure that [CLS] is prepended to your input sequence (see sample usage linked above)! Details for this model specifically:

• Architecture: gpt2
• Parameters: 124770816
• Maximum sequence length: 512 tokens
• Training text data (raw): 1335.62MB
• Training text data (byte premium scaled): 1000.005MB
• Training tokens: 274664448 (x10 epochs)
• Vocabulary size: 50000
• Compute cost: 1.401709989593088e+18 FLOPs or ~132.5 NVIDIA A6000 GPU hours

Training datasets (percentages prior to deduplication):

• 62.65815%: OSCAR 2021/09
• 37.34185%: Glot500, including CCNet, OSCAR, W2C, WikiMatrix

Citation

If you use this model, please cite:

@article{chang-etal-2024-goldfish,
  title={Goldfish: Monolingual Language Models for 350 Languages},
  author={Chang, Tyler A. and Arnett, Catherine and Tu, Zhuowen and Bergen, Benjamin K.},
  journal={Preprint},
  year={2024},
  url={https://www.arxiv.org/abs/2408.10441},
}