mmsegmentation导出onnx模型的问题

看到了导出的onnx自动执行了sofmax操作，不想要这个操作

修改如下代码：改为

在encoder_decoder_mapy.py

def simple_test(self, img, img_meta, rescale=True):
    """Simple test with single image."""
    # MOD
    # seg_logit = self.inference(img, img_meta, rescale)
    seg_logit, maps = self.inference(img, img_meta, rescale)
    if torch.onnx.is_in_onnx_export():
        # our inference backend only support 4D output
        seg_pred = seg_logit.unsqueeze(0)
        return seg_pred
    seg_pred = seg_logit.argmax(dim=1)
    seg_pred = seg_pred.cpu().numpy()
    # unravel batch dim
    seg_pred = list(seg_pred)
    # MOD
    # return seg_pred
    return seg_pred, maps

torch.onnx.is_in_onnx_export()是判断是否是onnx输出流，将argmax操作放到后面就可以了