goldfish-models commited on
Commit
7b0390e
1 Parent(s): d88a1dd

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +5 -3
README.md CHANGED
@@ -2,9 +2,9 @@
2
  ---
3
  license: apache-2.0
4
  language:
5
- - fas
6
  - prs
7
  - per
 
8
  datasets:
9
  - allenai/nllb
10
  - cis-lmu/Glot500
@@ -12,7 +12,7 @@ library_name: transformers
12
  pipeline_tag: text-generation
13
  tags:
14
  - goldfish
15
-
16
  ---
17
 
18
  # prs_arab_10mb
@@ -23,7 +23,7 @@ The Goldfish models are trained primarily for comparability across languages and
23
 
24
  Note: prs_arab is an [individual language](https://iso639-3.sil.org/code_tables/639/data) code. Macrolanguage code fas_arab (Persian) is included in Goldfish. Consider using that model depending on your use case.
25
 
26
- All training and hyperparameter details are in our paper, [Goldfish: Monolingual Language Models for 350 Languages (Chang et al., 2024)](https://github.com/tylerachang/goldfish/blob/main/goldfish_paper_20240815.pdf).
27
 
28
  Training code and sample usage: https://github.com/tylerachang/goldfish
29
 
@@ -33,6 +33,7 @@ Sample usage also in this Google Colab: [link](https://colab.research.google.com
33
 
34
  To access all Goldfish model details programmatically, see https://github.com/tylerachang/goldfish/blob/main/model_details.json.
35
  All models are trained with a [CLS] (same as [BOS]) token prepended, and a [SEP] (same as [EOS]) token separating sequences.
 
36
  Details for this model specifically:
37
 
38
  * Architecture: gpt2
@@ -59,5 +60,6 @@ If you use this model, please cite:
59
  author={Chang, Tyler A. and Arnett, Catherine and Tu, Zhuowen and Bergen, Benjamin K.},
60
  journal={Preprint},
61
  year={2024},
 
62
  }
63
  ```
 
2
  ---
3
  license: apache-2.0
4
  language:
 
5
  - prs
6
  - per
7
+ - fas
8
  datasets:
9
  - allenai/nllb
10
  - cis-lmu/Glot500
 
12
  pipeline_tag: text-generation
13
  tags:
14
  - goldfish
15
+ - arxiv:2408.10441
16
  ---
17
 
18
  # prs_arab_10mb
 
23
 
24
  Note: prs_arab is an [individual language](https://iso639-3.sil.org/code_tables/639/data) code. Macrolanguage code fas_arab (Persian) is included in Goldfish. Consider using that model depending on your use case.
25
 
26
+ 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).
27
 
28
  Training code and sample usage: https://github.com/tylerachang/goldfish
29
 
 
33
 
34
  To access all Goldfish model details programmatically, see https://github.com/tylerachang/goldfish/blob/main/model_details.json.
35
  All models are trained with a [CLS] (same as [BOS]) token prepended, and a [SEP] (same as [EOS]) token separating sequences.
36
+ For best results, make sure that [CLS] is prepended to your input sequence (see sample usage linked above)!
37
  Details for this model specifically:
38
 
39
  * Architecture: gpt2
 
60
  author={Chang, Tyler A. and Arnett, Catherine and Tu, Zhuowen and Bergen, Benjamin K.},
61
  journal={Preprint},
62
  year={2024},
63
+ url={https://www.arxiv.org/abs/2408.10441},
64
  }
65
  ```