## Dataset Summary This dataset is a **2 Billion token subset** of the [RedPajama-Data-1T](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T) dataset created by Together Computer. It was created to provide a smaller, lightweight version of the massive 1.2 Trillion token dataset for testing, debugging, and small-scale pre-training experiments. This specific subset was streamed primarily from the **C4 (Colossal Clean Crawled Corpus)** partition of the original RedPajama dataset. ## Dataset Structure The dataset retains the original structure of RedPajama. Each row contains: - **text**: The actual content (string). - **meta**: Metadata dict containing `url`, `timestamp`, `source`, etc. - **subset**: The name of the RedPajama subset this sample came from (e.g., `c4`). ## Source Data This dataset is a direct downstream derivative of **RedPajama-Data-1T**. - **Original Repository:** [togethercomputer/RedPajama-Data-1T](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T) - **Original Author:** Together Computer - **Subset Used:** `c4` (Colossal Clean Crawled Corpus) ## Creation Process This dataset was generated using the Hugging Face `datasets` library with the following methodology: 1. **Direct Streaming:** Data was streamed directly from the original source files to avoid downloading the full 5TB dataset. 2. **Selection:** The first ~2 Billion tokens were selected from the `c4` split. 3. **Token Counting:** Token counts were estimated using the `EleutherAI/gpt-neox-20b` tokenizer (or a 1 token ≈ 4 characters approximation). ## Licensing Information Since this is a subset of RedPajama, it inherits the licensing terms of the original data. - **RedPajama-Data-1T** is distributed under the licenses of its underlying data sources. - **C4 (Common Crawl)**: Terms of Use are available [here](https://commoncrawl.org/terms-of-use/). - **Code/Scripts**: The code used to generate this subset is Open Source. ## Citation If you use this dataset, please cite the original RedPajama paper: ```bibtex @software{together2023redpajama, author = {Together Computer}, title = {RedPajama: An Open Source Recipe to Reproduce LLaMA training dataset}, month = April, year = 2023, url = {[https://github.com/togethercomputer/RedPajama-Data](https://github.com/togethercomputer/RedPajama-Data)} }