Datasets:
Download data/train/HTML/part.4.parquet from andstor/the_pile_github: direct link, hf CLI and curl.
- Browser
- Download file 150 MB
-
https://huggingface.co/datasets/andstor/the_pile_github/resolve/main/data/train/HTML/part.4.parquet
- Command line
-
hf download hf://datasets/andstor/the_pile_github/data/train/HTML/part.4.parquet
-
curl -L -o part.4.parquet https://huggingface.co/datasets/andstor/the_pile_github/resolve/main/data/train/HTML/part.4.parquet
150 MB
- Xet hash:
- 1a16123f9f3f4e318f8b3cd0cdfaad0672fa39694316dc99ad4f8c0777e6c5fa
- Size of remote file:
- 150 MB
- SHA256:
- 746bcc0fed5e262288409578d4780b5f9f91adba7f743b09d5d6fa4377efe6e0
·
Xet efficiently stores Large Files inside Git, intelligently splitting files into unique chunks and accelerating uploads and downloads. More info.