Implement EasyCache and Invent LazyCache (#9496)

* Attempting a universal implementation of EasyCache, starting with flux as test; I screwed up the math a bit, but when I set it just right it works.

* Fixed math to make threshold work as expected, refactored code to use EasyCacheHolder instead of a dict wrapped by object

* Use sigmas from transformer_options instead of timesteps to be compatible with a greater amount of models, make end_percent work

* Make log statement when not skipping useful, preparing for per-cond caching

* Added DIFFUSION_MODEL wrapper around forward function for wan model

* Add subsampling for heuristic inputs

* Add subsampling to output_prev (output_prev_subsampled now)

* Properly consider conds in EasyCache logic

* Created SuperEasyCache to test what happens if caching and reuse is moved outside the scope of conds, added PREDICT_NOISE wrapper to facilitate this test

* Change max reuse_threshold to 3.0

* Mark EasyCache/SuperEasyCache as experimental (beta)

* Make Lumina2 compatible with EasyCache

* Add EasyCache support for Qwen Image

* Fix missing comma, curse you Cursor

* Add EasyCache support to AceStep

* Add EasyCache support to Chroma

* Added EasyCache support to Cosmos Predict t2i

* Make EasyCache not crash with Cosmos Predict ImagToVideo latents, but does not work well at all

* Add EasyCache support to hidream

* Added EasyCache support to hunyuan video

* Added EasyCache support to hunyuan3d

* Added EasyCache support to LTXV (not very good, but does not crash)

* Implemented EasyCache for aura_flow

* Renamed SuperEasyCache to LazyCache, hardcoded subsample_factor to 8 on nodes

* Eatra logging when verbose is true for EasyCache
This commit is contained in:
Jedrzej Kosinski
2025-08-22 19:41:08 -07:00
committed by GitHub
parent fe31ad0276
commit fc247150fe
17 changed files with 639 additions and 7 deletions

View File

@@ -1,6 +1,7 @@
#Based on Flux code because of weird hunyuan video code license.
import torch
import comfy.patcher_extension
import comfy.ldm.flux.layers
import comfy.ldm.modules.diffusionmodules.mmdit
from comfy.ldm.modules.attention import optimized_attention
@@ -348,6 +349,13 @@ class HunyuanVideo(nn.Module):
return repeat(img_ids, "t h w c -> b (t h w) c", b=bs)
def forward(self, x, timestep, context, y, guidance=None, attention_mask=None, guiding_frame_index=None, ref_latent=None, control=None, transformer_options={}, **kwargs):
return comfy.patcher_extension.WrapperExecutor.new_class_executor(
self._forward,
self,
comfy.patcher_extension.get_all_wrappers(comfy.patcher_extension.WrappersMP.DIFFUSION_MODEL, transformer_options)
).execute(x, timestep, context, y, guidance, attention_mask, guiding_frame_index, ref_latent, control, transformer_options, **kwargs)
def _forward(self, x, timestep, context, y, guidance=None, attention_mask=None, guiding_frame_index=None, ref_latent=None, control=None, transformer_options={}, **kwargs):
bs, c, t, h, w = x.shape
img_ids = self.img_ids(x)
txt_ids = torch.zeros((bs, context.shape[1], 3), device=x.device, dtype=x.dtype)