danbraunai/pile-uncopyrighted-tok download history

danbraunai/pile-uncopyrighted-tok is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 249 times (6 in the last 7 days), and 10,970 times in total. It ranks #56,764 among datasets by monthly downloads.

Pile Uncopyrighted (Tokenized) This is danbraunai/pile-uncopyrighted tokenized with EleutherAI/gpt-neox-20b. Each row contains a single input_ids column with 513 token IDs. Samples are concatenated with EOS tokens between them (following the approach in TransformerLens), then reshaped int

Open danbraunai/pile-uncopyrighted-tok on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.