onnxmodelzoo/inception-v1-12-int8

language: en license: apache-2.0 model_name: inception-v1-12-int8.onnx tags:

• validated
• vision
• classification
• inception_and_googlenet
• inception_v1

Inception v1

Compared with the fp32 Inception-1, int8 Inception-1's Top-1 accuracy drop ratio is -0.01% and performance improvement is 1.26x.

Note

The performance depends on the test hardware. Performance data here is collected with Intel® Xeon® Platinum 8280 Processor, 1s 4c per instance, CentOS Linux 8.3, data batch size is 1.

Description

Inception v1 is a reproduction of GoogLeNet.

Dataset

ILSVRC2012

Source

Caffe2 Inception v1 ==> ONNX Inception v1 ONNX Inception v1 ==> Quantized ONNX Inception v1

Model input and output

Input

data_0: float[1, 3, 224, 224]

Output

prob_1: float[1, 1000]

Pre-processing steps

Post-processing steps

Sample test data

random generated sampe test data:

• test_data_0.npz
• test_data_1.npz
• test_data_2.npz
• test_data_set_0
• test_data_set_1
• test_data_set_2

Results/accuracy on test set

Quantization

Inception-1-int8 and Inception-1-qdq are obtained by quantizing fp32 Inception-1 model. We use Intel® Neural Compressor with onnxruntime backend to perform quantization. View the instructions to understand how to use Intel® Neural Compressor for quantization.

Environment

onnx: 1.9.0 onnxruntime: 1.8.0

Prepare model

wget https://github.com/onnx/models/raw/main/vision/classification/inception_and_googlenet/inception_v1/model/inception-v1-12.onnx

Model quantize

Make sure to specify the appropriate dataset path in the configuration file.

bash run_tuning.sh --input_model=path/to/model \  # model path as *.onnx
--config=inception_v1.yaml \
--data_path=/path/to/imagenet \
--label_path=/path/to/imagenet/label \
--output_model=path/to/save

References

• Going deeper with convolutions

• Intel® Neural Compressor

Contributors

• mengniwang95 (Intel)
• airMeng (Intel)
• ftian1 (Intel)
• hshen14 (Intel)

License

MIT