The higgs dataset doesn't have headers, so we need to add `header=None` option to `pd.read_csv` call (https://github.com/NVIDIA/gbm-bench/blob/master/datasets.py#L172) to ensure that the first row isn't treated as header in pandas.
The higgs dataset doesn't have headers, so we need to add
header=Noneoption topd.read_csvcall (https://github.com/NVIDIA/gbm-bench/blob/master/datasets.py#L172) to ensure that the first row isn't treated as header in pandas.