theluyuan/DiffSynth-Studio
1
0
Fork 0
mirror of https://github.com/modelscope/DiffSynth-Studio.git synced 2026-03-19 06:39:43 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
d9fad821b21dd095a87a271196e47f8ce9fc31ac
DiffSynth-Studio/examples/TeaCache/README.md
Artiprocher c0889c2564 support teacache
2025-01-13 15:56:33 +08:00

1002 B
Raw Blame History

TeaCache

TeaCache (Timestep Embedding Aware Cache) is a training-free caching approach that estimates and leverages the fluctuating differences among model outputs across timesteps, thereby accelerating the inference.

Examples

We provide examples on FLUX.1-dev. See ./flux_teacache.py.

Steps: 50

GPU: A100

TeaCache is disabled tea_cache_l1_thresh=0.2 tea_cache_l1_thresh=0.4 tea_cache_l1_thresh=0.6 tea_cache_l1_thresh=0.8
23s 13s 9s 6s 5s
image_None image_0 2 image_0 4 image_0 6 image_0 8
Reference in New Issue View Git Blame Copy Permalink