For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Python module

max.pipelines.diffusion

Diffusion-specific pipeline components.

Pipelines

CompileWrapper	Wraps a compile target with optional input type annotations.
DiffusionPipeline	Base class for diffusion pipelines.
DiffusionPipelineOutput	Output of a diffusion pipeline.
PixelGenerationPipeline	Pixel generation pipeline for diffusion models.

First-block cache

FirstBlockCache	Standalone FirstBlockCache module.
FirstBlockCacheState	Per-request mutable state for FirstBlockCache.

Denoising cache

DenoisingCacheConfig	Pipeline-level cache configuration for diffusion model denoising.
DenoisingCacheState	Per-request mutable cache state for a single denoising stream.
TaylorSeer	Standalone TaylorSeer caching module.
TaylorSeerBufferState	Per-request mutable TaylorSeer state using Buffer objects.
TaylorSeerCache	High-level TaylorSeer for executor pipelines (Buffer-based).
TaylorSeerState	Per-request mutable state for TaylorSeer caching.

fbcache_conditional_execution	Handle FBCache F.cond branching pattern shared across DiT models.
max_compile	Decorator or function to compile a target with specified input types.
run_denoising_step	Execute one denoising step with caching logic.

Was this page helpful?

Thank you! We'll create more content like this.

Thank you for helping us improve!

😔 What went wrong?

Some code doesn’t work

It includes inaccurate information

It's missing information I need

It was difficult to understand

Other

Submit