| --- |
| license: other |
| language: |
| - code |
| task_categories: |
| - text-generation |
| pretty_name: DAMR Code 64K Tokenized |
| --- |
| |
| # DAMR Code 64K Tokenized |
|
|
| This gated repository contains 241,577,441,038 pretraining tokens derived from |
| [StarCoderData](https://huggingface.co/datasets/bigcode/starcoderdata). |
|
|
| ## Format |
|
|
| Files under `data/` contain contiguous little-endian `uint16` token IDs. |
| Concatenate shards in numeric order to reproduce the original stream. The |
| shared 64K BPE tokenizer is stored under `tokenizer/tokenizer.json`. Shard |
| manifests provide byte offsets, token counts, and SHA-256 checksums. |
|
|
| The intended production mixture uses Zyda2 at weight `0.858486`, this source at |
| `0.118510`, and math at `0.023004`. |
|
|
| ## Terms |
|
|
| Access and use are subject to the |
| [Terms of Use for The Stack](https://huggingface.co/datasets/bigcode/starcoderdata). |
| The source code is drawn from repositories with different licenses. Users must |
| comply with each original license, including attribution requirements where |
| applicable. Redistributed copies must include these terms and require |
| recipients to accept them. |
|
|
| This tokenized representation adds no ownership claim over the underlying |
| content. |
|
|
