LibraryActive

ONNX ModelContainer

API to store large tensors outside ModelProto, avoiding protobuf serialization of large initializers.

Field note

What it does

Implements an API to store large tensors outside the main ModelProto, avoiding copying large initializers when defining the model. No tensor is stored on disk until explicitly saved.

Capabilities

Available capabilities

Large tensor external storage DLPack protocol support Model serialization

ONNX ModelContainer

What it does

Available capabilities

Tags

Ways to use it

Product features

Helper functions to make ONNX graph components ¶

onnx.helper ¶

Other functions ¶

Tools ¶

Type Mappings ¶