goldfish-models
commited on
Commit
•
cd7028c
1
Parent(s):
f7d146b
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
@@ -2,8 +2,8 @@
|
|
2 |
---
|
3 |
license: apache-2.0
|
4 |
language:
|
5 |
-
- hbs
|
6 |
- srp
|
|
|
7 |
datasets:
|
8 |
- cis-lmu/Glot500
|
9 |
- oscar-corpus/OSCAR-2109
|
@@ -11,7 +11,7 @@ library_name: transformers
|
|
11 |
pipeline_tag: text-generation
|
12 |
tags:
|
13 |
- goldfish
|
14 |
-
|
15 |
---
|
16 |
|
17 |
# srp_latn_1000mb
|
@@ -24,7 +24,7 @@ Note: This language is available in Goldfish with other scripts (writing systems
|
|
24 |
|
25 |
Note: srp_latn 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 latn).
|
26 |
|
27 |
-
All training and hyperparameter details are in our paper, [Goldfish: Monolingual Language Models for 350 Languages (Chang et al., 2024)](https://
|
28 |
|
29 |
Training code and sample usage: https://github.com/tylerachang/goldfish
|
30 |
|
@@ -34,6 +34,7 @@ Sample usage also in this Google Colab: [link](https://colab.research.google.com
|
|
34 |
|
35 |
To access all Goldfish model details programmatically, see https://github.com/tylerachang/goldfish/blob/main/model_details.json.
|
36 |
All models are trained with a [CLS] (same as [BOS]) token prepended, and a [SEP] (same as [EOS]) token separating sequences.
|
|
|
37 |
Details for this model specifically:
|
38 |
|
39 |
* Architecture: gpt2
|
@@ -61,5 +62,6 @@ If you use this model, please cite:
|
|
61 |
author={Chang, Tyler A. and Arnett, Catherine and Tu, Zhuowen and Bergen, Benjamin K.},
|
62 |
journal={Preprint},
|
63 |
year={2024},
|
|
|
64 |
}
|
65 |
```
|
|
|
2 |
---
|
3 |
license: apache-2.0
|
4 |
language:
|
|
|
5 |
- srp
|
6 |
+
- hbs
|
7 |
datasets:
|
8 |
- cis-lmu/Glot500
|
9 |
- oscar-corpus/OSCAR-2109
|
|
|
11 |
pipeline_tag: text-generation
|
12 |
tags:
|
13 |
- goldfish
|
14 |
+
- arxiv:2408.10441
|
15 |
---
|
16 |
|
17 |
# srp_latn_1000mb
|
|
|
24 |
|
25 |
Note: srp_latn 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 latn).
|
26 |
|
27 |
+
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).
|
28 |
|
29 |
Training code and sample usage: https://github.com/tylerachang/goldfish
|
30 |
|
|
|
34 |
|
35 |
To access all Goldfish model details programmatically, see https://github.com/tylerachang/goldfish/blob/main/model_details.json.
|
36 |
All models are trained with a [CLS] (same as [BOS]) token prepended, and a [SEP] (same as [EOS]) token separating sequences.
|
37 |
+
For best results, make sure that [CLS] is prepended to your input sequence (see sample usage linked above)!
|
38 |
Details for this model specifically:
|
39 |
|
40 |
* Architecture: gpt2
|
|
|
62 |
author={Chang, Tyler A. and Arnett, Catherine and Tu, Zhuowen and Bergen, Benjamin K.},
|
63 |
journal={Preprint},
|
64 |
year={2024},
|
65 |
+
url={https://www.arxiv.org/abs/2408.10441},
|
66 |
}
|
67 |
```
|