-
Upload daozang_compact.db.gz with huggingface_hub
-
split train.parquet into 6 shards (<300MB each) so the HF dataset viewer works; README: shard layout + link DaoZang-Online space
-
single-split dataset: drop files.parquet (auto-inference requires standard split names and unified schema)
-
docs: embeddings are L2-normalized
-
add data/*.parquet (train: 285,117 chunks with bge-m3 embeddings; files: 3,152)
-
DaoZang dataset: 285,117 chunks with bge-m3 embeddings + 3,152 files + uncompressed markdowns
