|
--- |
|
task_categories: |
|
- text-generation |
|
language: |
|
- en |
|
pretty_name: RedPajama Tiny |
|
--- |
|
# Dataset Card for Dataset Name |
|
|
|
### Dataset Summary |
|
|
|
This is a tiny version of the RedPajama dataset, which is a clean-room, fully open-source implementation of the LLaMa dataset. |
|
|
|
This dataset contains 10 samples from each of the 7 sources. |
|
|
|
The full dataset has the following token counts and is available for [download]( https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T): |
|
|
|
| Dataset | Token Count | |
|
|---------------|-------------| |
|
| Commoncrawl | 878 Billion | |
|
| C4 | 175 Billion | |
|
| GitHub | 59 Billion | |
|
| Books | 26 Billion | |
|
| ArXiv | 28 Billion | |
|
| Wikipedia | 24 Billion | |
|
| StackExchange | 20 Billion | |
|
| Total | 1.2 Trillion | |
|
|
|
### Languages |
|
|
|
Primarily English, though the Wikipedia slice contains multiple languages. |
|
|
|
## Dataset Structure |
|
|
|
The dataset structure is as follows: |
|
|
|
``` |
|
{ |
|
"text": ..., |
|
"meta": {"url": "...", "timestamp": "...", "source": "...", "language": "...", ...} |
|
} |
|
``` |
|
|
|
## Dataset Creation |
|
|
|
This dataset was created to follow the LLaMa paper as closely as possible to try to reproduce its recipe. |
|
|
|
### Source Data |
|
|
|
#### Commoncrawl |
|
|
|
We download five dumps from Commoncrawl, and run the dumps through the official `cc_net` pipeline. |
|
We then deduplicate on the paragraph level, and filter out low quality text using a linear classifier trained to |
|
classify paragraphs as Wikipedia references or random Commoncrawl samples. |
|
|
|
#### C4 |
|
|
|
C4 is downloaded from Huggingface. The only preprocessing step is to bring the data into our own format. |
|
|
|
#### GitHub |
|
|
|
The raw GitHub data is downloaded from Google BigQuery. We deduplicate on the file level and filter out low quality |
|
files and only keep projects that are distributed under the MIT, BSD, or Apache license. |
|
|
|
#### Wikipedia |
|
We use the Wikipedia dataset available on Huggingface, which is based on the Wikipedia dump from 2023-03-20 and contains |
|
text in 20 different languages. The dataset comes in preprocessed format, so that hyperlinks, comments and other |
|
formatting boilerplate has been removed. |
|
|
|
#### Gutenberg and Books3 |
|
The PG19 subset of the Gutenberg Project and Books3 datasets are downloaded from Huggingface. After downloading, we use |
|
simhash to remove near duplicates. |
|
|
|
#### ArXiv |
|
ArXiv data is downloaded from Amazon S3 in the `arxiv` requester pays bucket. We only keep latex source files and |
|
remove preambles, comments, macros and bibliographies. |
|
|
|
#### Stackexchange |
|
The Stack Exchange split of the dataset is download from the |
|
[Internet Archive](https://archive.org/download/stackexchange). Here we only keep the posts from the 28 largest sites, |
|
remove html tags, group the posts into question-answer pairs, and order answers by their score. |
|
|
