Tensorflow模型中输入和输出名称之间的不一致

时间:2018-06-20 06:10:31

标签: tensorflow quantization

我已从Google网站下载了默认的optimized/quantized/stripped Inception-v3,并且具有以下节点:

Found 1 possible inputs: (name=Mul, type=float(1), shape=[])
No variables spotted.
Found 1 possible outputs: (name=softmax, op=Softmax)
Found 23834141 (23.83M) const parameters, 0 (0) variable parameters, and 0 control_edges
Op types used: 586 Const, 190 QuantizeDownAndShrinkRange, 95 QuantizedBiasAdd, 94 QuantizedConv2D, 94 QuantizedRelu, 11 QuantizedConcat, 9 QuantizedAvgPool, 5 QuantizedMaxPool, 3 Reshape, 2 Dequantize, 2 Max, 2 Min, 2 QuantizeV2, 1 Placeholder, 1 QuantizedMatMul, 1 Softmax

现在,这是我冻结并量化的预训练模型:

Found 1 possible inputs: (name=input, type=float(1), shape=[?,299,299,3])
No variables spotted.
Found 1 possible outputs: (name=InceptionV3/Predictions/Reshape_1, op=Dequantize)
Found 23824666 (23.82M) const parameters, 0 (0) variable parameters, and 0 control_edges
Op types used: 406 Const, 214 Requantize, 214 RequantizationRange, 134 Reshape, 134 Max, 134 Min, 134 QuantizeV2, 95 QuantizedConv2D, 94 QuantizedRelu, 94 QuantizedAdd, 49 Dequantize, 24 QuantizedMul, 15 ConcatV2, 10 QuantizedAvgPool, 4 QuantizedMaxPool, 2 QuantizedReshape, 1 QuantizedBiasAdd, 1 Placeholder, 1 Shape, 1 Softmax, 1 Squeeze

为什么输入节点名称和输出节点名称不一致?我尝试使用optimize_for_inference工具运行它,但没有解决任何问题。任何解释表示赞赏。

0 个答案:

没有答案