xddlj

> 尝试将模型从pytorch-》onnx-》ncnn，转换期间由于各框架支持算子的限制，将adaptiveAvepooling换成普通的pooling,F.interplate的插值方式从bilinear换成了nearest,经过训练在python上能得到非常好的分割结果，但是转换到ncnn进行C++推理后，同样的图片分割结果很差，请问有人转PSPnet碰到过这钟问题吗？ 请问一下，如何将adaptiveAvepooling换成普通的pooling,F.interplate的插值方式从bilinear换成了nearest,经过训练得到非常好的分割结果的？

> I have converted the model to onnx success: `torch.onnx.export(model, img, './test.onnx', verbose=True, opset_version=opset_version, input_names=input_names, output_names=output_names, dynamic_axes=dynamic_axes) ` When I export pt to onnx, has this error, Can you tell...

> Using opset11 RuntimeError: Exporting the operator deform_conv2d to ONNX opset version 11 is not supported. Please open a bug to request ONNX export support for the missing operator

+1 Could you please relese the trained models and a simply script to get the visualization result ?

+1 Can you provide a simply script to get the visualization result？ thanks you！