Implementation:Kornia Kornia ONNXLoader Load Model

Knowledge Sources	Kornia
Domains	ONNX, Deployment
Last Updated	2026-02-09 15:00 GMT

Overview

Concrete tool for loading ONNX models from various sources provided by Kornia's ONNX utility module.

Description

ONNXLoader.load_model is a static method that loads ONNX model files from local paths, URLs, or HuggingFace Hub. It uses CachedDownloader for network sources, supports loading external weight data, and returns onnx.ModelProto objects. HuggingFace Hub models use the "hf://operators/model_name" URI format.

Usage

Import and call as a static method before constructing ONNXSequential pipelines.

Code Reference

Repository	https://github.com/kornia/kornia
File	`kornia/onnx/utils.py`
Lines	L37–130
Signature	`class ONNXLoader: @staticmethod def load_model(model_name: str, download: bool = True, with_data: bool = False, **kwargs) -> onnx.ModelProto`
Import	`from kornia.onnx import ONNXLoader`

I/O Contract

Inputs

Parameter	Type	Required	Description
`model_name`	`str`	Yes	Model identifier — local path, URL, or `"hf://operators/name"`
`download`	`bool`	No	Download if not cached (default `True`)
`with_data`	`bool`	No	Load external weight data (default `False`)

Outputs

onnx.ModelProto — the loaded model object.

Usage Examples

Load from HuggingFace Hub

from kornia.onnx import ONNXLoader

model = ONNXLoader.load_model("hf://operators/resize")
print(type(model))  # <class 'onnx.onnx_ml_pb2.ModelProto'>

Load from local path

from kornia.onnx import ONNXLoader

model = ONNXLoader.load_model("/path/to/model.onnx")

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment