---
license: mit
size_categories:
- 10B<n<100B
---
`sample-10BT` version of the [FineWeb dataset](https://huggingface.co/datasets/HuggingFaceFW/fineweb) tokenized using the gpt2 tokenizer and split into 100M tokens binary shards.

A shard is simply a 1D stream of `np.uint16` numbers which are the tokenized samples from the dataset, stored contiguously.
Each sample from the dataset was prefixed with the `<|endoftext|>` special token before being tokenized.

There are 103 training shards (under `train/` dir) and 1 shard for validation (under `val/`).