Dataset · Hugging Face
clankur/longcrawl64
LongCrawl64 is a dataset for research on architectures and algorithms for long-context modeling. It consists of 6,661,465 pre-tokenized documents, each of which is 65,536 tokens long, for a total token count of 435 billion. The dataset is preprocessed with truncation to exactly 64 KiT, shuffling along document dimension, and rolling each document randomly along sequence dimension.
| Author / organization | clankur |
|---|---|
| License (as declared) | mit |
| Downloads | 944 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 22 January 2025 |
| Last updated at source | 19 February 2025 |
Popularity, last 2 days
944 now (-5 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
More from clankur
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.