share benchmark results in the docs with the dataset viewer
write outputs to csv
use
upload_file()
to upload benchmarks here (set
repo_type="dataset"
)
use the
README.md
in each dataset repo to configure the splits (see
here
for example)