Datasets:
Download default/train/0000.parquet from jblitzar/github-python-corpus: direct link, hf CLI and curl.
- Browser
- Download file 179 MB
-
https://huggingface.co/datasets/jblitzar/github-python-corpus/resolve/refs%2Fconvert%2Fparquet/default/train/0000.parquet
- Command line
-
hf download hf://datasets/jblitzar/github-python-corpus@refs/convert/parquet/default/train/0000.parquet
-
curl -L -o 0000.parquet https://huggingface.co/datasets/jblitzar/github-python-corpus/resolve/refs%2Fconvert%2Fparquet/default/train/0000.parquet
179 MB
- Xet hash:
- 43279160514584545208b7040e8d5d7f2d3daea7f43b5fbc4ac3fa3942957ee2
- Size of remote file:
- 179 MB
- SHA256:
- f24274e1bc31eb5d4158b5937b2c92a6f2dd3897107b6351227399d84d6f7626
·
Xet efficiently stores Large Files inside Git, intelligently splitting files into unique chunks and accelerating uploads and downloads. More info.