The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
OGBench, mirrored
A mirror of the OGBench dataset release by Park et al. It is not a new dataset and should not be
cited as one; the original lives at https://github.com/seohongpark/ogbench and the ogbench package
downloads it for you. This copy exists only because one of these directories is 39 GB and pulling it
repeatedly is slow.
manifest.json lists sha256 and byte size for every file. It is worth checking against: code that
loads a dataset by globbing a directory will silently pick up any stray .npz left in it, and an
extra file joins the training rotation without any error or log line.
Shard counts differ by directory on purpose. cube-quadruple-play-100m-v0 carries all 100 shards of
the 100M release; cube-triple-play-100m-v0 and puzzle-4x4-play-100m-v0 carry ten, the 10M subset.
Six further domains ship as a single .npz pair at the top level.
53.7 GB, 272 files.
- Downloads last month
- 347