官术网_书友最值得收藏!

Model inspection internals

On model inspection in the debugger, the following attributes can be found before calling the compile method:

input=Tensor("dense_1_input:0", shape=(?, 8), dtype=float32)
input_names=<class 'list'>: ['dense_1_input']
input_shape=<class 'tuple'>: (None, 8)
inputs=<class 'list'>: [<tf.Tensor 'dense_1_input:0' shape=(?, 8) dtype=float32>]
layers=<class 'list'>: [
<keras.layers.core.Dense object at 0x7fdbcbb444a8>,
<keras.layers.core.Dense object at 0x7fdbcbb05c50>,
<keras.layers.core.Dense object at 0x7fdbcbb05cf8>]
output=Tensor("dense_3/Sigmoid:0", shape=(?, 1), dtype=float32)
output_names=<class 'list'>: ['dense_3']
output_shape=<class 'tuple'>: (None, 1)
outputs-<class 'list'>: [<tf.Tensor 'dense_3/Sigmoid:0' shape=(?, 1) dtype=float32>]
trainable_weights=<class 'list'>:
[<tf.Variable 'dense_1/kernel:0' shape=(8, 12) dtype=float32_ref>,
<tf.Variable 'dense_1/bias:0' shape=(12,) dtype=float32_ref>,
<tf.Variable 'dense_2/kernel:0' shape=(12, 8) dtype=float32_ref>,
<tf.Variable 'dense_2/bias:0' shape=(8,) dtype=float32_ref>,
<tf.Variable 'dense_3/kernel:0' shape=(8, 1) dtype=float32_ref>,
<tf.Variable 'dense_3/bias:0' shape=(1,) dtype=float32_ref>]
weights=<class 'list'>:
[<tf.Variable 'dense_1/kernel:0' shape=(8, 12) dtype=float32_ref>,
<tf.Variable 'dense_1/bias:0' shape=(12,) dtype=float32_ref>,
<tf.Variable 'dense_2/kernel:0' shape=(12, 8) dtype=float32_ref>,
<tf.Variable 'dense_2/bias:0' shape=(8,) dtype=float32_ref>,
<tf.Variable 'dense_3/kernel:0' shape=(8, 1) dtype=float32_ref>,
<tf.Variable 'dense_3/bias:0' shape=(1,) dtype=float32_ref>]
主站蜘蛛池模板: 乌鲁木齐市| 同心县| 锦州市| 鲜城| 禹州市| 枣阳市| 西丰县| 尉犁县| 鄂托克前旗| 蒙自县| 泗水县| 娱乐| 东至县| 呼伦贝尔市| 拉萨市| 台北县| 堆龙德庆县| 威信县| 阳山县| 苏尼特右旗| 南城县| 额济纳旗| 大姚县| 岐山县| 亚东县| 唐海县| 霍邱县| 茶陵县| 手游| 孙吴县| 富宁县| 洱源县| 遵义市| 溆浦县| 永城市| 饶河县| 胶南市| 平昌县| 喀喇沁旗| 巴中市| 浏阳市|