File size: 2,862 Bytes
c03c978 b836885 c03c978 b836885 c03c978 b836885 c03c978 c1d2e6d c03c978 b836885 c03c978 b836885 c03c978 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 |
---
license: apache-2.0
language:
- zho
- lzh
- chi
datasets:
- cis-lmu/Glot500
- legacy-datasets/wikipedia
library_name: transformers
pipeline_tag: text-generation
tags:
- goldfish
- arxiv:2408.10441
---
# lzh_hant_full
Goldfish is a suite of monolingual language models trained for 350 languages.
This model is the <b>Literary Chinese</b> (Han Traditional script) model trained on 15MB of data (all our data in the language), after accounting for an estimated byte premium of 0.70; content-matched text in Literary Chinese takes on average 0.70x 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: lzh_hant is an [individual language](https://iso639-3.sil.org/code_tables/639/data) code. Macrolanguage code zho_hant (Chinese) is included in Goldfish. Consider using that model depending on your use case.
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): 10.64MB
* Training text data (byte premium scaled): 15.205MB
* Training tokens: 2767872 (x10 epochs)
* Vocabulary size: 50000
* Compute cost: 1.4119698825216e+16 FLOPs or ~1.3 NVIDIA A6000 GPU hours
Training datasets (percentages prior to deduplication):
* 53.82278%: [Wikipedia 2023/08](https://dumps.wikimedia.org/)
* 46.05529%: [Glot500](https://huggingface.co/datasets/cis-lmu/Glot500), including [Tatoeba](https://tatoeba.org/en/), [Wikipedia Hugging Face](https://huggingface.co/datasets/legacy-datasets/wikipedia)
* 0.12193%: [Tatoeba](https://tatoeba.org/en/)
## 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},
}
```
|