{
    "error_type": "shape_mismatch",
    "description": "输入tensor维度与模型期望不匹配",
    "error_patterns": [
        "shape mismatch: expected [1,3,224,224], got [1,3,256,256]",
        "dimension 1 mismatch: expected 224, got 256",
        "size mismatch in tensor 'input'",
        "tensor shape does not match model input",
        "input shape invalid for layer conv1"
    ],
    "diagnosis": {
        "cause": "输入数据的shape与模型转换时指定的shape不一致",
        "check_points": [
            "模型转换时是否指定了--inputShape参数",
            "实际输入数据的维度大小",
            "模型原始ONNX的input tensor定义"
        ]
    },
    "suggestions": [
        "检查输入数据的shape是否与模型定义一致",
        "使用benchmark的--inputShape参数指定正确维度",
        "重新转换模型时使用converter_lite的--inputShape参数固化shape"
    ],
    "example_log": "[ERROR] shape mismatch: input tensor shape [1,3,256,256] does not match model input shape [1,3,224,224]. Expected shape: [1,3,224,224], got: [1,3,256,256]"
}