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.graph

APIs to build inference graphs for MAX.

Graph construction

Graph	Represents a single MAX graph.
KernelLibrary	Manages custom kernel libraries and operations for a graph.

Graph values

BufferValue	Represents a mutable semantic tensor within a Graph.
BufferValueLike	Represent a PEP 604 union type
TensorValue	Represents a value semantic tensor within a Graph.
TensorValueLike	Represent a PEP 604 union type
Value	Represents a symbolic value within a Graph.

Type system

BufferType	A symbolic buffer type.
TensorType	A symbolic tensor type.
Type	The type of any value in a MAX graph.

Shapes and dimensions

AlgebraicDim	A dimension defined by an expression over symbolic dimensions.
Dim	A tensor dimension.
DimLike	Represent a PEP 604 union type
Shape	A sequence of Dim values representing a tensor shape.
ShapeLike	alias of Iterable[int	str	Dim	integer[Any]	TypedAttr]
StaticDim	A static tensor dimension with a fixed size.
SymbolicDim	A symbolic tensor dimension with an unknown size identified by name.

Devices

DeviceKind	A device type representation.
DevicePlacementPolicy	Controls behavior when an op implicitly transfers a tensor to CPU.
DeviceRef	A symbolic device representation.

Weights

ShardingStrategy	Specifies how a Weight should be sharded across multiple devices.
Weight	Represents a value in a Graph that can be loaded at a later time.

Configuration

GraphDebugConfig	Narrow view of max.engine.DebugConfig exposed through Graph.debug.

Submodules

• max.graph.ops
• max.graph.quantization
• max.graph.weights

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