test(frank-medium): A/B — flip to DFlash2 drafter (SPEC_METHOD=dflash, SPEC=7)

2026-09-20 22:06:26 -07:00
parent c1a4ccbbcd
commit 1ae73db084
1 changed files with 6 additions and 5 deletions
@@ -110,11 +110,12 @@ Environment=MAXLEN=262144
 # 400 here. Vision traffic routes to frank-small on dankclaw (harness
 # modelRoles vision -> frank-small).
 Environment=GPU_UTIL=0.98
-# Speculative decoding: MTP self-speculation at depth 8 (the model's own
-# single-layer MTP head, applied autoregressively). The documented plain-FP8
-# path (upstream docker-compose) uses num_speculative_tokens=8.
-Environment=SPEC_METHOD=mtp
-Environment=SPEC=8
+# Speculative decoding: A/B test — DFlash2 external drafter at depth 7 (the
+# documented DFlash2 depth), to compare against the MTP self-speculation
+# baseline (depth 8, mean accept 5.66/8). The DFlash2 drafter is a separate
+# quantization-agnostic checkpoint at /models/Qwen3.8-27B-DFlash2-FP8.
+Environment=SPEC_METHOD=dflash
+Environment=SPEC=7
 # KV cache: 0 = profile the shape (the documented plain-FP8 path's default;
 # the upstream docker-compose leaves the pin commented out). The measured pin
 # for this shape is 15989735424 (14.89 GiB, upstream, 2x R9700 32GB); after