Skip to content

Dataset · HuggingFaceBio

1.1 trillion nucleotide DNA and RNA genomic corpus

HuggingFaceBio/carbon-pretraining-corpus

173 million eukaryotic and prokaryotic DNA and RNA sequences for pretraining base genomic models.

What's inside

A large-scale biological corpus in parquet format containing 173 million DNA and RNA sequences from various eukaryotic and prokaryotic species. Across four main configurations, the volume reaches 1.1 trillion base pairs (about 180 billion tokens for a 6-mer tokenizer). The data was prepared by HuggingFaceBio for pretraining the Carbon foundation genomic model.

What you can do with it

  • Pretraining base genomic language models
  • Studying DNA and RNA sequences of various organisms
  • Benchmarking bioinformatics algorithms and tokenizers
Tasks
Text generation
Size
100M–1B rows
Downloads this month
2.9K
Likes
34

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("HuggingFaceBio/carbon-pretraining-corpus")