Merge nucleic/sleek-ember-seal-uady into dev

2026-07-31 00:00:13 -07:00
parent ac71c67e9a
commit 0bee416a88
1 changed files with 9 additions and 0 deletions
@@ -274,6 +274,15 @@ MLX remains Metal-first. `--device cpu` is an explicit diagnostic fallback for p
 checks and bounded smoke tests; it is not an acceptable full-training path when Metal is
 available.

+The completed Metal run stopped after epoch two and selected epoch one at 94.42%
+fake-quant validation accuracy and a 95.59% teacher-aware selection score. Its selected
+float checkpoint improves frozen v1 to **95.09% (891/937)**, two correct decisions above
+the accepted baseline's float checkpoint. That gain does not survive export: the matched
+23,148,500-byte int8-QDQ graph scores **94.66% (887/937)**, 94.66% macro recall, and
+94.17% scored-hard accuracy. It is five correct decisions behind the accepted int8
+baseline and fails the 95% shipping gate. Preserve the artifact as a rejected experiment;
+`purpose-lite-v1-distilled-qat-mlx-4e` remains the candidate of record.
+
 ### Convert and validate Core ML

 Core ML Tools no longer maintains the legacy ONNX converter, so the Apple artifact is