Download:
- release 1.1: https://s3.amazonaws.com/download.onnx/models/opset_3/resnet50.tar.gz
- release 1.1.2: https://s3.amazonaws.com/download.onnx/models/opset_6/resnet50.tar.gz
- release 1.2: https://s3.amazonaws.com/download.onnx/models/opset_7/resnet50.tar.gz
- release 1.3: https://s3.amazonaws.com/download.onnx/models/opset_8/resnet50.tar.gz
- master: https://s3.amazonaws.com/download.onnx/models/opset_9/resnet50.tar.gz
Model size: 103 MB
ResNet-50 is a deep convolutional networks for classification.
Deep Residual Learning for Image Recognition
Run ResNet-50 in browser - implemented by ONNX.js with ResNet-50 release 1.2
Caffe2 ResNet-50 ==> ONNX ResNet-50
- Mean: 128
- Std: 128
- Scale: 256
- Reference: the training script
gpu_0/data_0: float[1, 3, 224, 224]
gpu_0/softmax_1: float[1, 1000]
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
MIT