An error occurs when running the onnx model

[yang162132]

An error occurs when running the onnx model

2024-01-09 11:06:18.1575181 [E:onnxruntime:, sequential_executor.cc:346 onnxruntime::SequentialExecutor::Execute] Non-zero status code returned while running ReduceMax node. Name:'ReduceMax_1745' Status Message: D:\a\_work\1\s\onnxruntime\core\providers\cpu\reduction\reduction_ops.cc:747 onnxruntime::ValidateKeepDims keepdims was false. Can't reduce on dim with value of 0 if 'keepdims' is false. Invalid output shape would be produced. input_shape:{0,4}

this onnx file is converted from the pytorch pth file

    device = torch.device("cpu")
    model = create_model(num_classes=2)
    train_weights = "./save_weights/model.pth"
    model.load_state_dict(torch.load(train_weights, map_location=device)["model"])
    model.to(device)
    model.eval()
    original_img = Image.open("G:/yangben/newdata/png/56.png")
    data_transform = transforms.Compose([transforms.ToTensor()])
    img = data_transform(original_img)
    data_transform = transforms.Compose([transforms.ToTensor()])
    img = data_transform(original_img)
    img = torch.unsqueeze(img, dim=0)
    torch.onnx.export(model, img,
                      "faster_rcnn.onnx", 
                      opset_version=11, 
    )

[zachgk]

Where are you getting this error? If I am reading it right, there is a call to max() in your model that is not supported by ONNX (but is presumably supported by Pytorch). You may have to modify the model to not make that call. It may also have something to do with a dimension (which may be the batch size) that has a value of 0 and that is not supported

[yang162132]

Where are you getting this error? If I am reading it right, there is a call to max() in your model that is not supported by ONNX (but is presumably supported by Pytorch). You may have to modify the model to not make that call. It may also have something to do with a dimension (which may be the batch size) that has a value of 0 and that is not supported

Do you know where I need to adjust in my model? What function is this problem generally related to?