IMPORTANT: To view this page as Markdown, append `.md` to the URL (e.g. /max/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. /max/get-started.md).

Python class

Partial

`Partial`

class max.experimental.sharding.Partial(reduce_op=ReduceOp.SUM)

source

Bases: Placement

Every device holds a partial result that must be reduced.

Parameters:: reduce_op (ReduceOp) – The reduction operation to apply. Defaults to ReduceOp.SUM.

`reduce_op`

reduce_op: ReduceOp = 'sum'

source

`transition_to()`

transition_to(other)

source

Partial-to-Replicated is allreduce; Partial-to-Sharded is reduce-scatter.

Parameters:: other (Placement)
Return type:: Collective

Partial
- reduce_op
- transition_to()