Add the 80 benchmark sequences (manifest.csv/.parquet) and dataset card

#3

The 80 collagen triple-helix target sequences of the CDSM benchmark, one row per RCSB PDB entry (30 homotrimers, 50 heterotrimers), as CSV and parquet, with a dataset card documenting the schema, loading, provenance, licensing and citation.

mjbuehler changed pull request status to merged

Sign up or log in to comment