feat(campaign): protected-cast applier, recipe markers, KLD gate runner

apply-protected-cast.sh splices derive-protected-experts.py output into the recipe's PROTECTED-BEGIN/END section (refuses to re-derive, trashes the pre-edit recipe). kld-gate-swift15.sh boots the candidate on the serving pin in --kld-ref mode (pool fully resident, no disk tier) and enforces mean KL<=0.03 / top-1>=95% from the JSON report.
2026-10-06 22:41:18 -07:00
parent 07af1a9366
commit fa858f4c10
3 changed files with 131 additions and 0 deletions
@@ -30,6 +30,7 @@
 # law -- an expert whose share of its layer's down-input energy is at least 3% and at least three
 # times its share of the routing, an even count on each rank's half of the layer (the four-bit GEMM
 # takes the rest as two equal tables by parity) by adding that half's next most energetic expert.
+# PROTECTED-BEGIN (rewritten from Swift's OWN grams by derive-protected-experts.py)
 blk.34.ffn_*_exps.407.weight         cast dtype=bf16
 blk.34.ffn_*_exps.496.weight         cast dtype=bf16
 blk.44.ffn_*_exps.292.weight         cast dtype=bf16
@@ -40,6 +41,7 @@ blk.47.ffn_*_exps.122.weight         cast dtype=bf16
 blk.47.ffn_*_exps.143.weight         cast dtype=bf16
 blk.47.ffn_*_exps.399.weight         cast dtype=bf16
 blk.47.ffn_*_exps.445.weight         cast dtype=bf16
+# PROTECTED-END

 # Every other routed expert, trunk and MTP layer alike: the down projection as E4M3 codes with a
 # bf16 scale a 128x128 block -- one amax per block, scale = amax/448 rounded to bf16, the exact