如何从 tflite 模型中的输出细节确定边界框、类别和分数的索引

如何解决如何从 tflite 模型中的输出细节确定边界框、类别和分数的索引

这些是解释器打印的输出细节

[{'name': 'BoxPredictor_0/BoxEncodingPredictor/BiasAdd','index': 0,'shape': array([ 1,20,12]),'shape_signature': array([ 1,'dtype': <class 'numpy.uint8'>,'quantization': (0.0235294122248888,0),'quantization_parameters': {'scales': array([0.02352941],dtype=float32),'zero_points': array([0]),'quantized_dimension': 0},'sparsity_parameters': {}},{'name': 'BoxPredictor_0/Classpredictor/BiasAdd','index': 1,'shape': array([  1,273]),'shape_signature': array([  1,{'name': 'BoxPredictor_1/BoxEncodingPredictor/BiasAdd','index': 2,10,24]),{'name': 'BoxPredictor_1/Classpredictor/BiasAdd','index': 3,546]),{'name': 'BoxPredictor_2/BoxEncodingPredictor/BiasAdd','index': 4,5,{'name': 'BoxPredictor_2/Classpredictor/BiasAdd','index': 5,{'name': 'BoxPredictor_3/BoxEncodingPredictor/BiasAdd','index': 6,3,{'name': 'BoxPredictor_3/Classpredictor/BiasAdd','index': 7,{'name': 'BoxPredictor_4/BoxEncodingPredictor/BiasAdd','index': 8,2,{'name': 'BoxPredictor_4/Classpredictor/BiasAdd','index': 9,{'name': 'BoxPredictor_5/BoxEncodingPredictor/BiasAdd','index': 10,1,{'name': 'BoxPredictor_5/Classpredictor/BiasAdd','index': 11,'sparsity_parameters': {}}]