CoolFace
Datasetpublic

avewright/chess-soft-sf19

avewright/chess-soft-sf19 Official Stockfish 19 MultiPV soft targets. This release supersedes the 25k pilot. It is not a filter of chess-soft-multipv-lichess or chess-soft-100m-disagreements. 2,010,006 rows. Source id 4. Vocab compact (1968). Mix (as generated) origin rows note self-play (origin=1) 0 SF19 vs SF19, ε=0.20, book + 4 random legal relabel (origin=0) 0 existing local boards, new SF19 labels frozen eval 10,000 split=1 in… See the full description on the dataset page: https://huggingface.co/datasets/avewright/chess-soft-sf19.

sourceHugging Facemitupdated 13d agoView on Hugging Face
0likes628downloads
Dataset Card

avewright/chess-soft-sf19

Official Stockfish 19 MultiPV soft targets. This release supersedes the 25k pilot. It is not a filter of chess-soft-multipv-lichess or chess-soft-100m-disagreements.

2,010,006 rows. Source id 4. Vocab compact (1968).

Mix (as generated)

originrowsnote
self-play (origin=1)0SF19 vs SF19, ε=0.20, book + 4 random legal
relabel (origin=0)0existing local boards, new SF19 labels
frozen eval10,000split=1 in data/shard_000000.parquet

Relabel was intended to be 80%. The existing-board iterator stopped at ~84k accepted labels; self-play filled the 1M train remainder. split=0 is train.

Phase (train+eval new rows): opening 0 / middlegame 0 / endgame 0. Eval bucket: equal 0 / winning 0 / losing 0 / mate 0.

Teacher

  • —Stockfish 19 tag sf_19 (edb0d9db), EvalFile nn-1a298aa575a0.nnue
  • —Full file SHA-256 in teacher.json
  • —Full strength, Threads=1, Hash=32, UCI_ShowWDL=true
  • —Ponder left to python-chess (do not set it via configure)
  • —Hash is not cleared between production searches
  • —Label budget: 100k nodes / MultiPV=8 / tau=120
  • —Play (self-play only): 4k-node cheap search on unlabeled plies, label every 3 plies after ply 4

Targets

  • —Policy: STM softmax(tau=120) over the last complete MultiPV-8 iteration
  • —Unsearched legal moves are absent, not proven bad
  • —Mate rank is sign * (100000 - min(|mate|, 1000)), not mate-as-cp
  • —Bound scores are dropped
  • —cp / mate / wdl: White-absolute (training loader contract)
  • —soft_indices / soft_probs: width 8, pad -1 / 0
  • —soft_cps / soft_mates are stored so softmax can be rebuilt without SF
  • —FEN-only relabel rows drop repetition history and may omit EP if no legal capture

Honor split. shard_000000 is the frozen eval set (saved_split_v1). Do not invent a new position-hash holdout.

Quality audit (2,000 isolated positions, 100k/8 vs 1M/8)

  • —top-1 vs 1M: None (self-play ?, relabel ?; endgame worst)
  • —regret p50=0 / p90=0 / mean=0
  • —missing 1M-ref mass p50=0.0% / p90=0.0%

See audit.json. Flags did not predict disagreement; no adaptive extra search.

Files

  • —data/shard_XXXXXX.parquet — one inbox shard per file
  • —teacher.json, summary.json, sampling.json, audit.json, eval_manifest.json