ONNX Models

This is a repository for storing ONNX models.

Models

• BVLC AlexNet (217 MByte)

• BVLC GoogleNet (26 MByte)

• BVLC CaffeNet (217 MByte)

• BVLC R-CNN ILSVRC13 (206 MByte)

• DenseNet-121 (31 MByte)

• Inception-v1 (26 MByte)

• Inception-v2 (41 MByte)

• ResNet-50 (92 MByte)

• ShuffleNet (6.7 MByte)

• SqueezeNet (6.0 MByte)

• VGG-16 (310 MByte)

• VGG-19 (510 MByte)

• MNIST (26 kByte)

• Emotion-FERPlus (34 MByte)

Usage

Every ONNX backend should support running these models out of the box. After dowloading and extracting the tarball of each model, there should be

• A protobuf file model.onnx which is the serialized ONNX model.
• Several sets of sample inputs and outputs files (test_data_*.npz), they are numpy serialized archive.

e.g. they can be used like this:

import numpy as np
import onnx
import onnx_backend as backend

# Load the model and sample inputs and outputs
model = onnx.load(model_pb_path)
sample = np.load(npz_path, encoding='bytes')
inputs = list(sample['inputs'])
outputs = list(sample['outputs'])

# Run the model with an onnx backend and verify the results
np.testing.assert_almost_equal(outputs, backend.run_model(model, inputs))

License

MIT License