|
|
|
--- |
|
license: apache-2.0 |
|
language: |
|
- orm |
|
datasets: |
|
- cis-lmu/Glot500 |
|
- castorini/afriberta-corpus |
|
- statmt/cc100 |
|
- legacy-datasets/wikipedia |
|
- csebuetnlp/xlsum |
|
- allenai/nllb |
|
- allenai/MADLAD-400 |
|
library_name: transformers |
|
pipeline_tag: text-generation |
|
tags: |
|
- goldfish |
|
- arxiv:2408.10441 |
|
--- |
|
|
|
# orm_latn_100mb |
|
|
|
Goldfish is a suite of monolingual language models trained for 350 languages. |
|
This model is the <b>Oromo</b> (Latin script) model trained on 100MB of data, after accounting for an estimated byte premium of 1.26; content-matched text in Oromo takes on average 1.26x 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: orm_latn is a [macrolanguage](https://iso639-3.sil.org/code_tables/639/data) code. Individual language code gaz_latn (West Central Oromo) 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)](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): 126.44MB |
|
* Training text data (byte premium scaled): 100.005MB |
|
* Training tokens: 28832256 (x10 epochs) |
|
* Vocabulary size: 50000 |
|
* Compute cost: 1.47118126989312e+17 FLOPs or ~13.9 NVIDIA A6000 GPU hours |
|
|
|
Training datasets (percentages prior to deduplication): |
|
* 32.67081%: [Glot500](https://huggingface.co/datasets/cis-lmu/Glot500), including [AfriBERTa](https://huggingface.co/datasets/castorini/afriberta-corpus), [AfroMAFT](https://zenodo.org/record/6990611#.Y0-yU-xBw-Q), [CC100](https://huggingface.co/datasets/statmt/cc100), [CCNet](https://github.com/facebookresearch/cc_net), [HornMT](https://github.com/asmelashteka/HornMT), [Wortschatz Leipzig Data](https://wortschatz.uni-leipzig.de/en/download), [MoT](https://github.com/bltlab/mot), [Parallel Corpora for Ethiopian Languages](https://github.com/AAUThematic4LT/Parallel-Corpora-for-Ethiopian-Languages), [TICO](https://tico-19.github.io/), [Wikipedia Hugging Face](https://huggingface.co/datasets/legacy-datasets/wikipedia), [XLSum](https://huggingface.co/datasets/csebuetnlp/xlsum) |
|
* 27.67571%: [NLLB (CommonCrawl and ParaCrawl)](https://huggingface.co/datasets/allenai/nllb) |
|
* 23.09955%: [MADLAD-400 (CommonCrawl)](https://huggingface.co/datasets/allenai/MADLAD-400) |
|
* 14.69511%: [AfriBERTa](https://huggingface.co/datasets/castorini/afriberta-corpus) |
|
* 1.12785%: [eBible](https://ebible.org/find/) |
|
* 0.73097%: [Wikipedia 2023/08](https://dumps.wikimedia.org/) |
|
|
|
|
|
## 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}, |
|
} |
|
``` |
|
|
