Hugging Face
Models
Datasets
Spaces
Posts
Docs
Solutions
Pricing
Log In
Sign Up
Datasets:
stanford-crfm
/
DSIR-filtered-pile-50M
like
8
Tasks:
Text Generation
Fill-Mask
Modalities:
Text
Formats:
json
Languages:
English
Size:
1M - 10M
ArXiv:
arxiv:
2302.03169
Tags:
language modeling
masked language modeling
pretraining
pile
DSIR
Libraries:
Datasets
Dask
Croissant
+ 1
License:
mit
Dataset card
Viewer
Files
Files and versions
Community
refs/convert/parquet
DSIR-filtered-pile-50M
/
default
/
partial-train
Commit History
Delete old duckdb index files
03dba6a
verified
parquet-converter
commited on
Mar 7
Update duckdb index files
fb32996
parquet-converter
commited on
Jan 7
Update parquet files
68bcdc3
parquet-converter
commited on
Sep 4, 2023