IMPORTANT: To view this page as Markdown, append `.md` to the URL (e.g. /get-started.md). For the complete documentation index, see llms.txt.
Skip to main content
For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Python class

SpecDecodeTarget

SpecDecodeTarget​

class max.pipelines.speculative.spec_target.SpecDecodeTarget(*args, **kwargs)

source

Bases: Protocol[_BatchT]

The target model’s entry point and output layout.

ep_input_types()​

ep_input_types()

source

Expert-parallel input types to append to the graph signature.

Return type:

Sequence[TensorType | BufferType]

verify()​

verify(batch)

source

Runs the target over the merged prompt + draft tokens.

Parameters:

batch (_BatchT)

Return type:

Verified