Cache methods speedup diffusion transformers by storing and reusing intermediate outputs of specific layers, such as attention and feedforward layers, instead of recalculating them at each inference step.

[[autodoc]] CacheMixin

[[autodoc]] PyramidAttentionBroadcastConfig

[[autodoc]] apply_pyramid_attention_broadcast

[[autodoc]] FasterCacheConfig

[[autodoc]] apply_faster_cache

[[autodoc]] FirstBlockCacheConfig

[[autodoc]] apply_first_block_cache