mirror of
https://github.com/modelscope/DiffSynth-Studio.git
synced 2026-03-18 22:08:13 +00:00
1002 B
1002 B
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 |