Area (Language):

大陆港澳 (简体中文)

台灣 (繁體中文)

 

 

[Introduction]

 

Model conversion, from yolo PyTorh model to ONNX model, or Tensorflow 1.X model to ONNX model.

 

 

[Operation steps and instructions]

 

• PyTroch to ONNX

 

 

Select the PyTorch model you want to convert, you can enter the file name yourself, or press select to select the model.

 

Note:

1. The PyTorch model to be converted must be placed in the model/pytorch folder.

2. The currently supported pytorch model conversion is yolov4’s pytorch model conversion to onnx, and yolov4 pytorch must be a darknet architecture.

3. The extension of the selected file only supports .pth files.

 

 

You can enter the batch size of the converted onnx model yourself. If you do not fill in, the default value is 1.

 

 

You can enter the converted onnx model file name yourself.

Note:

The converted onnx model file will be stored in the onnx folder in the model folder.

 

 

Press “convert” to convert the model. The converted model will be placed in the model/onnx folder.

 

 

 

• Tensorflow to ONNX

 

 

Select the Tensorflow model to be converted, you can enter the file name yourself, or click “Choose” to select the model.

 

 

1.The converted Tensorflow model must be placed in tensorflow in the model folder.

2.The currently supported Tensorflow model is downloaded from https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf1_detection_zoo.md, which can only support the model trained by tensorflow1.X.

3.The extension of the selected file only supports .pb files.

 

 

You can enter the converted onnx model file name yourself.

Note:

The converted onnx model file will be stored in the onnx folder in the model folder.