Commit
·
d68a67a
1
Parent(s):
5b580f2
Update README.md
Browse files
README.md
CHANGED
@@ -36,7 +36,7 @@ dataset_info:
|
|
36 |
|
37 |
### Dataset Summary
|
38 |
|
39 |
-
This dataset is one of the two datasets published by
|
40 |
The dataset available at this link is the pre-training data constrained by vocabulary. The other published data i.e. the pre-training data that is not constrained by vocabulary is available at https://huggingface.co/datasets/text-machine-lab/unconstrained_language.
|
41 |
|
42 |
The vocabulary used for curating the data is constructed from the AOChildes corpus (https://www.sciencedirect.com/science/article/abs/pii/S0079742121000256). The AOChildes corpus consists of transcripts of child-directed speech. Hence, the vocabulary constructed from AOChildes corpus consists of words spoken or heard by children of age six years or younger.
|
@@ -46,13 +46,13 @@ The vocabulary is then used to filter the widely used text corpora,
|
|
46 |
- BookCorpus: https://ieeexplore.ieee.org/document/7410368,
|
47 |
- Wikipedia: https://huggingface.co/datasets/wikipedia,
|
48 |
- Simplified-Wikipedia: https://simple.wikipedia.org/wiki/Main_Page,
|
49 |
-
-
|
50 |
|
51 |
-
From above corpora, only those spans are included
|
52 |
|
53 |
### Languages
|
54 |
|
55 |
-
The dataset contains English language only.
|
56 |
|
57 |
## Dataset Structure
|
58 |
|
@@ -64,7 +64,7 @@ Every data instance has only one key "Text" that included a text span of approxi
|
|
64 |
|
65 |
### Citation Information
|
66 |
|
67 |
-
```
|
68 |
@article{deshpande2023honey,
|
69 |
title={Honey, I Shrunk the Language: Language Model Behavior at Reduced Scale},
|
70 |
author={Deshpande, Vijeta and Pechi, Dan and Thatte, Shree and Lialin, Vladislav and Rumshisky, Anna},
|
|
|
36 |
|
37 |
### Dataset Summary
|
38 |
|
39 |
+
This dataset is one of the two datasets published by "Honey, I Shrunk the Language: Language Model Behavior at Reduced Scale".
|
40 |
The dataset available at this link is the pre-training data constrained by vocabulary. The other published data i.e. the pre-training data that is not constrained by vocabulary is available at https://huggingface.co/datasets/text-machine-lab/unconstrained_language.
|
41 |
|
42 |
The vocabulary used for curating the data is constructed from the AOChildes corpus (https://www.sciencedirect.com/science/article/abs/pii/S0079742121000256). The AOChildes corpus consists of transcripts of child-directed speech. Hence, the vocabulary constructed from AOChildes corpus consists of words spoken or heard by children of age six years or younger.
|
|
|
46 |
- BookCorpus: https://ieeexplore.ieee.org/document/7410368,
|
47 |
- Wikipedia: https://huggingface.co/datasets/wikipedia,
|
48 |
- Simplified-Wikipedia: https://simple.wikipedia.org/wiki/Main_Page,
|
49 |
+
- Children's Book Test Corpus: https://arxiv.org/abs/1511.02301
|
50 |
|
51 |
+
From the above corpora, only those spans are included that contain words only from the predefined vocabulary. The dataset includes 44 million sentences (~6 million sequences of ~128 tokens) and 3 million contiguous spans with approximately 128 tokens.
|
52 |
|
53 |
### Languages
|
54 |
|
55 |
+
The dataset contains the English language only.
|
56 |
|
57 |
## Dataset Structure
|
58 |
|
|
|
64 |
|
65 |
### Citation Information
|
66 |
|
67 |
+
```sh
|
68 |
@article{deshpande2023honey,
|
69 |
title={Honey, I Shrunk the Language: Language Model Behavior at Reduced Scale},
|
70 |
author={Deshpande, Vijeta and Pechi, Dan and Thatte, Shree and Lialin, Vladislav and Rumshisky, Anna},
|