71 lines
3.4 KiB
Markdown
71 lines
3.4 KiB
Markdown
|
|
---
|
|
license: apache-2.0
|
|
language:
|
|
- tat
|
|
datasets:
|
|
- allenai/MADLAD-400
|
|
- cis-lmu/Glot500
|
|
- legacy-datasets/wikipedia
|
|
- allenai/nllb
|
|
- oscar-corpus/OSCAR-2109
|
|
library_name: transformers
|
|
pipeline_tag: text-generation
|
|
tags:
|
|
- goldfish
|
|
- arxiv:2408.10441
|
|
---
|
|
|
|
# tat_cyrl_1000mb
|
|
|
|
Goldfish is a suite of monolingual language models trained for 350 languages.
|
|
This model is the <b>Tatar</b> (Cyrillic script) model trained on 1000MB of data, after accounting for an estimated byte premium of 1.85; content-matched text in Tatar takes on average 1.85x 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: tat_cyrl is an [individual language](https://iso639-3.sil.org/code_tables/639/data) code. It is not contained in any macrolanguage codes contained in Goldfish (for script cyrl).
|
|
|
|
All training and hyperparameter details are in our paper, [Goldfish: Monolingual Language Models for 350 Languages (Chang et al., 2024)](https://www.arxiv.org/abs/2408.10441).
|
|
|
|
Training code and sample usage: https://github.com/tylerachang/goldfish
|
|
|
|
Sample usage also in this Google Colab: [link](https://colab.research.google.com/drive/1rHFpnQsyXJ32ONwCosWZ7frjOYjbGCXG?usp=sharing)
|
|
|
|
## 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): 1854.54MB
|
|
* Training text data (byte premium scaled): 1000.005MB
|
|
* Training tokens: 232933888 (x10 epochs)
|
|
* Vocabulary size: 50000
|
|
* Compute cost: 1.18882910011392e+18 FLOPs or ~112.4 NVIDIA A6000 GPU hours
|
|
|
|
Training datasets (percentages prior to deduplication):
|
|
* 42.95952%: [MADLAD-400 (CommonCrawl)](https://huggingface.co/datasets/allenai/MADLAD-400)
|
|
* 28.38275%: [Glot500](https://huggingface.co/datasets/cis-lmu/Glot500), including [Earthlings](https://publicdata.canterbury.ac.nz/Research/Geocorpus/CCGLU_v5.0/), [Wortschatz Leipzig Data](https://wortschatz.uni-leipzig.de/en/download), [OSCAR](https://oscar-project.org/), [TIL](https://github.com/turkic-interlingua/til-mt), [W2C](https://lindat.mff.cuni.cz/repository/xmlui/handle/11858/00-097C-0000-0022-6133-9), [Wikipedia Hugging Face](https://huggingface.co/datasets/legacy-datasets/wikipedia), [WikiMatrix](https://github.com/facebookresearch/LASER/tree/main/tasks/WikiMatrix)
|
|
* 12.13842%: [NLLB (CommonCrawl and ParaCrawl)](https://huggingface.co/datasets/allenai/nllb)
|
|
* 10.01257%: [OSCAR 2021/09](https://huggingface.co/datasets/oscar-corpus/OSCAR-2109)
|
|
* 4.69512%: [Wikipedia 2023/08](https://dumps.wikimedia.org/)
|
|
* 1.81162%: [Languages of Russia](http://web-corpora.net/wsgi3/minorlangs/download)
|
|
|
|
|
|
## 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},
|
|
}
|
|
```
|