This PR improves the TabBioMed results dataset card by adding the task categories `tabular-classification` and `tabular-regression`, reflecting the benchmark’s classification, regression, and multi-target tabular scope.
It also adds the curated tags `medical` and `biology`, since the benchmark covers biomedical tabular data including electronic health records, drug discovery, genomics, transcriptomics, proteomics, metabolomics, single-cell omics, and systems biology. The existing `benchmark`, `tabular`, and `biomedical` tags are preserved. The `synthetic` tag was considered but skipped because the benchmark datasets are curated from public sources.
Finally, links to the code repository and live leaderboard are added. No license metadata is added because the README states that the Apache-2.0 code license is separate from dataset and model-checkpoint licenses.