fix: reuse conditioning cache for reference-free Fixed V3.9 runs

Reuse the existing bounded prompt-conditioning cache only for safe reference-free Fixed groups. Preserve Reference, List/Timeline, terminal-pair, Sampling and Run Storage paths; add 22 regression cases and bilingual update guidance.

Local validation: 1558 CPU passed, 1 skipped; eight successful GPU generations. Post-fix same-session warm total medians: V3.8 273.579s vs V3.9 273.779s, with overlapping ranges. This scoped result is not a universal speed or GPU-kernel claim. Official workflows/ZIPs and historical Releases are unchanged.
This commit is contained in:
ukr8b3g-cmyk
2026-10-01 16:13:48 +09:00
parent 29e0ce3dd3
commit 21a857115d
8 changed files with 316 additions and 74 deletions
+19 -1
View File
@@ -25,7 +25,7 @@ from ..reference import (
prepare_reference_assets,
resolve_reference_image_inputs,
)
from ..v2.h3_builder import encode_prompt_conditioning
from ..v2.h3_builder import encode_prompt_conditioning, encode_prompt_conditioning_cached
from .reference_effective_plan import PictureMap, compile_group_picture_map
from .reference_routing import REFERENCE_SLOT_IDS, ReferenceRoutingSchedule
@@ -243,6 +243,8 @@ class ReferenceRoutingRuntime:
timeline_video_assets: Any = None,
reference_encode_cache: bool = False,
cache_event: Any = None,
prompt_conditioning_cache: bool = False,
prompt_cache_event: Any = None,
expected_group_contract: dict[str, Any] | None = None,
first_frame_hash: str = "none",
last_frame_hash: str = "none",
@@ -315,6 +317,22 @@ class ReferenceRoutingRuntime:
reference_audio_assets=reference_audio_assets,
timeline_video_assets=timeline_video_assets,
)
elif (prompt_conditioning_cache and not self.has_selected_references
and reference_audio_assets is None and timeline_video_assets is None
and terminal_prompt_policy is None):
# The Sequence caller enables this only for Fixed, reference-free
# runs. Reuse the V3.8 cache and its conservative CLIP guards; keep
# route/storage validation above and group metadata below intact.
conditioning = encode_prompt_conditioning_cached(
clip, effective_prompt,
first_image=first_image,
last_image=presented_last_image,
first_image_fingerprint=first_frame_hash,
last_image_fingerprint=(last_frame_hash
if presented_last_image is not None else "none"),
cache_enabled=True,
cache_event=prompt_cache_event,
)
else:
conditioning = encode_prompt_conditioning(
clip, effective_prompt,