ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Reflow verbose diffusion comments to fewer lines

Browse source
This commit is contained in:
Daniel Han 2026-07-13 03:30:48 +00:00
commit a3388820a2
28 changed files with 136 additions and 184 deletions
Download patch file Download diff file Expand all files Collapse all files

49
studio/backend/core/inference/diffusion_cache.py
View file

@ -34,13 +34,12 @@ TC_MODES = (TC_FBCACHE, TC_MAGCACHE)
DEFAULT_FBCACHE_THRESHOLD = 0.08
QUANT_FBCACHE_THRESHOLD = 0.12
# MagCache (diffusers >= 0.39): skips whole steps from a PRE-CALIBRATED residual-magnitude
# curve with an accumulated-error budget, a skip cap, and an early-step no-skip retention
# window -- so unlike FBCache the divergence from the uncached trajectory is bounded.
# Measured (HunyuanVideo-1.5-720p, B200, 50 steps): threshold 0.12 = 1.5x e2e at LPIPS
# 0.147 vs the same uncached stack with the SAME composition (FBCache at its 0.08 default
# reached 2.4x but LPIPS 0.54: a brighter, different clip -- why the fbcache auto policy
# excludes this family).
# MagCache (diffusers >= 0.39): skips whole steps from a PRE-CALIBRATED residual-magnitude curve
# with an accumulated-error budget, a skip cap, and an early-step no-skip retention window -- so
# unlike FBCache the divergence from the uncached trajectory is bounded. Measured
# (HunyuanVideo-1.5-720p, B200, 50 steps): threshold 0.12 = 1.5x e2e at LPIPS 0.147 vs the same
# uncached stack with the SAME composition (FBCache at its 0.08 default reached 2.4x but LPIPS 0.54:
# a brighter, different clip -- why the fbcache auto policy excludes this family).
DEFAULT_MAGCACHE_THRESHOLD = 0.12
MAGCACHE_MAX_SKIP_STEPS = 3
MAGCACHE_RETENTION_RATIO = 0.2
@ -302,13 +301,12 @@ _MAGCACHE_WAN5B_RATIOS = (
0.9208,
)
# All curves are calibrated at the 50-step schedule, so a single-DiT curve has 50 entries
# and MagCacheConfig interpolates it to the actual step count. A dual-expert MoE
# (Wan2.2-A14B) runs each expert on a SLICE of the schedule and the hook counts each
# expert's OWN forwards from 0, so each expert carries its own curve (keyed
# "family::transformer_2" for the second), sized to that expert's steps in the 50-step
# calibration; engage-time scales it by the requested step count (the boundary split is a
# fixed fraction of the schedule).
# All curves are calibrated at the 50-step schedule, so a single-DiT curve has 50 entries and
# MagCacheConfig interpolates it to the actual step count. A dual-expert MoE (Wan2.2-A14B) runs each
# expert on a SLICE of the schedule and the hook counts each expert's OWN forwards from 0, so each
# expert carries its own curve (keyed "family::transformer_2" for the second), sized to that
# expert's steps in the 50-step calibration; engage-time scales it by the requested step count (the
# boundary split is a fixed fraction of the schedule).
_MAGCACHE_CALIBRATION_STEPS = 50
_MAGCACHE_FAMILY_RATIOS: dict[str, tuple[float, ...]] = {
@ -328,18 +326,17 @@ def _magcache_ratio_key(family: Optional[str], expert: Optional[str]) -> str:
return f"{fam}::{exp}"
# Families whose AUTO step-cache decision engages MagCache instead of FBCache. On
# HunyuanVideo-1.5 FBCache free-runs and derails the trajectory (LPIPS 0.54 + a luma shift
# at its default), while MagCache holds the same composition at 1.5x. On Wan2.2-TI2V-5B
# both stay composition-true but MagCache dominates (B200, 1280x704/33f/50 steps, pairwise
# LPIPS): balanced MagCache 1.65x/0.034 vs FBCache 0.08 at 1.49x/0.031, fast points
# 1.73x/0.044 vs 1.71x/0.083. On Wan2.2-A14B (dual-expert MoE) the OPPOSITE holds (B200,
# 1280x720/33f/50 steps, per-expert curves): FBCache 0.12 at 2.88x/0.128 dominates balanced
# MagCache (1.80x/0.145), and FBCache 0.08 at 1.28x/0.098 beats MagCache quality's
# 1.14x/0.074 -- the 16-step high-noise expert leaves MagCache too few forwards to skip
# within its budget -- so it keeps FBCache and ships no curve (an explicit magcache request
# runs uncached with a warning). Every other family keeps FBCache. An EXPLICIT
# "fbcache"/"magcache" request always wins.
# Families whose AUTO step-cache decision engages MagCache instead of FBCache. On HunyuanVideo-1.5
# FBCache free-runs and derails the trajectory (LPIPS 0.54 + a luma shift at its default), while
# MagCache holds the same composition at 1.5x. On Wan2.2-TI2V-5B both stay composition-true but
# MagCache dominates (B200, 1280x704/33f/50 steps, pairwise LPIPS): balanced MagCache 1.65x/0.034 vs
# FBCache 0.08 at 1.49x/0.031, fast points 1.73x/0.044 vs 1.71x/0.083. On Wan2.2-A14B (dual-expert
# MoE) the OPPOSITE holds (B200, 1280x720/33f/50 steps, per-expert curves): FBCache 0.12 at
# 2.88x/0.128 dominates balanced MagCache (1.80x/0.145), and FBCache 0.08 at 1.28x/0.098 beats
# MagCache quality's 1.14x/0.074 -- the 16-step high-noise expert leaves MagCache too few forwards
# to skip within its budget -- so it keeps FBCache and ships no curve (an explicit magcache request
# runs uncached with a warning). Every other family keeps FBCache. An EXPLICIT "fbcache"/"magcache"
# request always wins.
_FAMILY_AUTO_CACHE_MODE: dict[str, str] = {
"hunyuanvideo-1.5": TC_MAGCACHE,
"hunyuanvideo-1.5-720p": TC_MAGCACHE,