Dataset Viewer
Auto-converted to Parquet Duplicate
text
stringclasses
5 values
cheeky-cornflower-setter-02e496ce4abb-20220421-092639
cheeky-cornflower-setter-02e496ce4abb-20220421-093149
cheeky-cornflower-setter-032f4eb01e68-20220418-125807
cheeky-cornflower-setter-054834d1b04a-20220419-154302
cheeky-cornflower-setter-0a5ba522405b-20220422-133010

vrisingwm-mc-data

Training data for mc-neurosymbolic-ledger (code: github.com/wyzardWu/mc-neurosymbolic-ledger).

Backup status (2026-07-19, final)

vpt_mc_352/ is a complete mirror of the training set, uploaded before the dev machine was returned: 1953 label files (seg_000.npz) + 1878 latent files (seg_000_lat44.npy) — byte-identical to the local copy (75 segments never had latents; regenerable via ingest/vpt6xx_ingest.py from public VPT).

The top-level <player>/seg_000_latent.npy folders are the older v3-era dataset (different format), kept for the v3 baseline.

Note: a backup-completion note for the private repo README is committed locally on the dev machine but unpushed (GitHub auth expired overnight); git push origin main from the machine sends it, otherwise this file is the authoritative record.

Downloads last month
1,496