opening expert model card

README.md

@@ -18,8 +18,8 @@ finetuned on `≥26`-piece positions (one-hot best first move) from
 [`Lichess/chess-position-evaluations`](https://huggingface.co/datasets/Lichess/chess-position-evaluations)
 via [`avewright/lichess-opening-bestline`](https://huggingface.co/datasets/avewright/lichess-opening-bestline).
 
-This file is **`latest.pt` at opening-FT step 750** (2026-09-14 12:41 UTC).
-Train loss ~1.8435. Frozen holdout hard CE ~1.5031.
+This file is **`latest.pt` at opening-FT step 900** (2026-09-14 12:44 UTC).
+Train loss ~1.8286. Frozen holdout hard CE ~1.5039.
 
 Not the generalist incumbent, the puzzle expert, the Syzygy expert, or the endgame expert.
 
@@ -28,12 +28,12 @@ Not the generalist incumbent, the puzzle expert, the Syzygy expert, or the endga
 - Warm start: public 99M `latest.pt` (weights only), then full resume.
 - Split: position-hash 80/20 (seed 277). Frozen piece-stratified val 8192.
 - One-hot PV1 (`soft_alpha=0`). Pieces 26–32.
-- Polar-NorMuon, bs=528. Best disk ckpt at upload: step 750.
+- Polar-NorMuon, bs=528. Best disk ckpt at upload: step 900.
 
 ## Files
 
 - `latest.pt`
-- `step_000750.pt`
+- `step_000900.pt`
 - `model_config.json`
 - `train.log`
 - `pack.json`