TensorFlow网络结构重新定义

最新推荐文章于 2022-12-11 22:22:11 发布
最新推荐文章于 2022-12-11 22:22:11 发布
阅读量1.5k 收藏
点赞数 2
分类专栏： speechdnn 文章标签： TensorFlow 网络结构
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/SPESEG/article/details/100551889
版权

今天我想把网络结构重新定义下，如图下


这种定义的弊端就是太高大上，call根本没有必要，而且用了tf网络模板，这是装逼。

本人想把这种去掉，取而代之的是简单易懂的结构。

不妨关注我博客，随时更新，人生不易，请你珍惜眼前的我。

 

 

 

 

 

 

 

 

 

 

 

请

充

值

查

看

完

整

版

，

我

要

积

分

！

！

！

 

 

 

 

 

 

 

 

 

 

 

补充一：我直接把这些玩意去掉了，模型也可以训练，奇怪的是模型训练次数越多，模型越大，而变大的是meta，data一直没有变化，所以关键在于将浮点转为更小的，这样更易于应用落地。

补充二：我特么发现错误，FailedPreconditionError (see above for traceback): Attempting to use uninitialized value

我恢复模型初始化了啊，这是什么鬼？？？仅仅简单修改了下网络定义而已。。。。

补充三：我先不管测试了，我特么转pb。但是我明明定义了节点名字啊，为啥还是啥都没有？？据我所知，上面的节点根本不是节点名字，我用另外一种方法得到的节点是

x_mixed
global_step/initial_value
global_step
global_step/Assign
global_step/read
save/filename/input
save/filename
save/Const
save/SaveV2/tensor_names
save/SaveV2/shape_and_slices
save/SaveV2
save/control_dependency
save/RestoreV2/tensor_names
save/RestoreV2/shape_and_slices
save/RestoreV2
save/Assign
save/restore_all

我擦，这是啥玩意，节点呢？？？？？？？？？

补充四：卧槽，老子创造性思维就是厉害，加模型进来就有了节点名字，之前是没有加进来模型的缘故，

顺利转成pb，下面就看调用pb是否成功了，人生真是峰回路转啊。

节点名字：

x_mixed
global_step/initial_value
global_step
global_step/Assign
global_step/read
rnn/Rank
rnn/range/start
rnn/range/delta
rnn/range
rnn/concat/values_0
rnn/concat/axis
rnn/concat
rnn/transpose
rnn/Shape
rnn/strided_slice/stack
rnn/strided_slice/stack_1
rnn/strided_slice/stack_2
rnn/strided_slice
rnn/MultiRNNCellZeroState/GRUCellZeroState/ExpandDims/dim
rnn/MultiRNNCellZeroState/GRUCellZeroState/ExpandDims
rnn/MultiRNNCellZeroState/GRUCellZeroState/Const
rnn/MultiRNNCellZeroState/GRUCellZeroState/concat/axis
rnn/MultiRNNCellZeroState/GRUCellZeroState/concat
rnn/MultiRNNCellZeroState/GRUCellZeroState/zeros/Const
rnn/MultiRNNCellZeroState/GRUCellZeroState/zeros
rnn/MultiRNNCellZeroState/GRUCellZeroState/ExpandDims_1/dim
rnn/MultiRNNCellZeroState/GRUCellZeroState/ExpandDims_1
rnn/MultiRNNCellZeroState/GRUCellZeroState/Const_1
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/ExpandDims/dim
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/ExpandDims
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/Const
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/concat/axis
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/concat
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/zeros/Const
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/zeros
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/ExpandDims_1/dim
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/ExpandDims_1
rnn/MultiRNNCellZeroState/GRUCellZeroState_1/Const_1
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/ExpandDims/dim
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/ExpandDims
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/Const
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/concat/axis
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/concat
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/zeros/Const
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/zeros
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/ExpandDims_1/dim
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/ExpandDims_1
rnn/MultiRNNCellZeroState/GRUCellZeroState_2/Const_1
rnn/Shape_1
rnn/strided_slice_1/stack
rnn/strided_slice_1/stack_1
rnn/strided_slice_1/stack_2
rnn/strided_slice_1
rnn/Shape_2
rnn/strided_slice_2/stack
rnn/strided_slice_2/stack_1
rnn/strided_slice_2/stack_2
rnn/strided_slice_2
rnn/ExpandDims/dim
rnn/ExpandDims
rnn/Const
rnn/concat_1/axis
rnn/concat_1
rnn/zeros/Const
rnn/zeros
rnn/time
rnn/TensorArray
rnn/TensorArray_1
rnn/TensorArrayUnstack/Shape
rnn/TensorArrayUnstack/strided_slice/stack
rnn/TensorArrayUnstack/strided_slice/stack_1
rnn/TensorArrayUnstack/strided_slice/stack_2
rnn/TensorArrayUnstack/strided_slice
rnn/TensorArrayUnstack/range/start
rnn/TensorArrayUnstack/range/delta
rnn/TensorArrayUnstack/range
rnn/TensorArrayUnstack/TensorArrayScatter/TensorArrayScatterV3
rnn/Maximum/x
rnn/Maximum
rnn/Minimum
rnn/while/iteration_counter
rnn/while/Enter
rnn/while/Enter_1
rnn/while/Enter_2
rnn/while/Enter_3
rnn/while/Enter_4
rnn/while/Enter_5
rnn/while/Merge
rnn/while/Merge_1
rnn/while/Merge_2
rnn/while/Merge_3
rnn/while/Merge_4
rnn/while/Merge_5
rnn/while/Less
rnn/while/Less/Enter
rnn/while/Less_1
rnn/while/Less_1/Enter
rnn/while/LogicalAnd
rnn/while/LoopCond
rnn/while/Switch
rnn/while/Switch_1
rnn/while/Switch_2
rnn/while/Switch_3
rnn/while/Switch_4
rnn/while/Switch_5
rnn/while/Identity
rnn/while/Identity_1
rnn/while/Identity_2
rnn/while/Identity_3
rnn/while/Identity_4
rnn/while/Identity_5
rnn/while/add/y
rnn/while/add
rnn/while/TensorArrayReadV3
rnn/while/TensorArrayReadV3/Enter
rnn/while/TensorArrayReadV3/Enter_1
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/Initializer/random_uniform/shape
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/Initializer/random_uniform/min
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/Initializer/random_uniform/max
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/Initializer/random_uniform/RandomUniform
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/Initializer/random_uniform/sub
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/Initializer/random_uniform/mul
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/Initializer/random_uniform
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/Assign
rnn/multi_rnn_cell/cell_0/gru_cell/gates/kernel/read
rnn/multi_rnn_cell/cell_0/gru_cell/gates/bias/Initializer/Const
rnn/multi_rnn_cell/cell_0/gru_cell/gates/bias
rnn/multi_rnn_cell/cell_0/gru_cell/gates/bias/Assign
rnn/multi_rnn_cell/cell_0/gru_cell/gates/bias/read
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/Initializer/random_uniform/shape
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/Initializer/random_uniform/min
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/Initializer/random_uniform/max
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/Initializer/random_uniform/RandomUniform
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/Initializer/random_uniform/sub
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/Initializer/random_uniform/mul
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/Initializer/random_uniform
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/Assign
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/kernel/read
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/bias/Initializer/zeros
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/bias
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/bias/Assign
rnn/multi_rnn_cell/cell_0/gru_cell/candidate/bias/read
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/concat/axis
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/concat
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/MatMul
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/MatMul/Enter
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/BiasAdd
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/BiasAdd/Enter
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/Sigmoid
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/Const
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/split/split_dim
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/split
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/mul
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/concat_1/axis
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/concat_1
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/MatMul_1
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/MatMul_1/Enter
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/BiasAdd_1
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/BiasAdd_1/Enter
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/Tanh
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/mul_1
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/sub/x
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/sub
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/mul_2
rnn/while/rnn/multi_rnn_cell/cell_0/gru_cell/add
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/Initializer/random_uniform/shape
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/Initializer/random_uniform/min
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/Initializer/random_uniform/max
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/Initializer/random_uniform/RandomUniform
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/Initializer/random_uniform/sub
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/Initializer/random_uniform/mul
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/Initializer/random_uniform
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/Assign
rnn/multi_rnn_cell/cell_1/gru_cell/gates/kernel/read
rnn/multi_rnn_cell/cell_1/gru_cell/gates/bias/Initializer/Const
rnn/multi_rnn_cell/cell_1/gru_cell/gates/bias
rnn/multi_rnn_cell/cell_1/gru_cell/gates/bias/Assign
rnn/multi_rnn_cell/cell_1/gru_cell/gates/bias/read
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/Initializer/random_uniform/shape
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/Initializer/random_uniform/min
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/Initializer/random_uniform/max
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/Initializer/random_uniform/RandomUniform
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/Initializer/random_uniform/sub
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/Initializer/random_uniform/mul
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/Initializer/random_uniform
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/Assign
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/kernel/read
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/bias/Initializer/zeros
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/bias
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/bias/Assign
rnn/multi_rnn_cell/cell_1/gru_cell/candidate/bias/read
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/concat/axis
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/concat
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/MatMul
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/MatMul/Enter
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/BiasAdd
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/BiasAdd/Enter
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/Sigmoid
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/Const
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/split/split_dim
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/split
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/mul
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/concat_1/axis
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/concat_1
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/MatMul_1
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/MatMul_1/Enter
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/BiasAdd_1
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/BiasAdd_1/Enter
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/Tanh
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/mul_1
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/sub/x
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/sub
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/mul_2
rnn/while/rnn/multi_rnn_cell/cell_1/gru_cell/add
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/Initializer/random_uniform/shape
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/Initializer/random_uniform/min
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/Initializer/random_uniform/max
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/Initializer/random_uniform/RandomUniform
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/Initializer/random_uniform/sub
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/Initializer/random_uniform/mul
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/Initializer/random_uniform
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/Assign
rnn/multi_rnn_cell/cell_2/gru_cell/gates/kernel/read
rnn/multi_rnn_cell/cell_2/gru_cell/gates/bias/Initializer/Const
rnn/multi_rnn_cell/cell_2/gru_cell/gates/bias
rnn/multi_rnn_cell/cell_2/gru_cell/gates/bias/Assign
rnn/multi_rnn_cell/cell_2/gru_cell/gates/bias/read
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/Initializer/random_uniform/shape
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/Initializer/random_uniform/min
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/Initializer/random_uniform/max
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/Initializer/random_uniform/RandomUniform
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/Initializer/random_uniform/sub
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/Initializer/random_uniform/mul
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/Initializer/random_uniform
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/Assign
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/kernel/read
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/bias/Initializer/zeros
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/bias
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/bias/Assign
rnn/multi_rnn_cell/cell_2/gru_cell/candidate/bias/read
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/concat/axis
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/concat
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/MatMul
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/MatMul/Enter
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/BiasAdd
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/BiasAdd/Enter
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/Sigmoid
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/Const
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/split/split_dim
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/split
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/mul
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/concat_1/axis
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/concat_1
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/MatMul_1
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/MatMul_1/Enter
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/BiasAdd_1
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/BiasAdd_1/Enter
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/Tanh
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/mul_1
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/sub/x
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/sub
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/mul_2
rnn/while/rnn/multi_rnn_cell/cell_2/gru_cell/add
rnn/while/TensorArrayWrite/TensorArrayWriteV3
rnn/while/TensorArrayWrite/TensorArrayWriteV3/Enter
rnn/while/add_1/y
rnn/while/add_1
rnn/while/NextIteration
rnn/while/NextIteration_1
rnn/while/NextIteration_2
rnn/while/NextIteration_3
rnn/while/NextIteration_4
rnn/while/NextIteration_5
rnn/while/Exit
rnn/while/Exit_1
rnn/while/Exit_2
rnn/while/Exit_3
rnn/while/Exit_4
rnn/while/Exit_5
rnn/TensorArrayStack/TensorArraySizeV3
rnn/TensorArrayStack/range/start
rnn/TensorArrayStack/range/delta
rnn/TensorArrayStack/range
rnn/TensorArrayStack/TensorArrayGatherV3
rnn/Const_1
rnn/Rank_1
rnn/range_1/start
rnn/range_1/delta
rnn/range_1
rnn/concat_2/values_0
rnn/concat_2/axis
rnn/concat_2
rnn/transpose_1
y_hat_src1/kernel/Initializer/random_uniform/shape
y_hat_src1/kernel/Initializer/random_uniform/min
y_hat_src1/kernel/Initializer/random_uniform/max
y_hat_src1/kernel/Initializer/random_uniform/RandomUniform
y_hat_src1/kernel/Initializer/random_uniform/sub
y_hat_src1/kernel/Initializer/random_uniform/mul
y_hat_src1/kernel/Initializer/random_uniform
y_hat_src1/kernel
y_hat_src1/kernel/Assign
y_hat_src1/kernel/read
y_hat_src1/bias/Initializer/zeros
y_hat_src1/bias
y_hat_src1/bias/Assign
y_hat_src1/bias/read
y_hat_src1/Tensordot/axes
y_hat_src1/Tensordot/free
y_hat_src1/Tensordot/Shape
y_hat_src1/Tensordot/GatherV2/axis
y_hat_src1/Tensordot/GatherV2
y_hat_src1/Tensordot/GatherV2_1/axis
y_hat_src1/Tensordot/GatherV2_1
y_hat_src1/Tensordot/Const
y_hat_src1/Tensordot/Prod
y_hat_src1/Tensordot/Const_1
y_hat_src1/Tensordot/Prod_1
y_hat_src1/Tensordot/concat/axis
y_hat_src1/Tensordot/concat
y_hat_src1/Tensordot/stack
y_hat_src1/Tensordot/transpose
y_hat_src1/Tensordot/Reshape
y_hat_src1/Tensordot/transpose_1/perm
y_hat_src1/Tensordot/transpose_1
y_hat_src1/Tensordot/Reshape_1/shape
y_hat_src1/Tensordot/Reshape_1
y_hat_src1/Tensordot/MatMul
y_hat_src1/Tensordot/Const_2
y_hat_src1/Tensordot/concat_1/axis
y_hat_src1/Tensordot/concat_1
y_hat_src1/Tensordot
y_hat_src1/BiasAdd
y_hat_src1/Relu
y_hat_src2/kernel/Initializer/random_uniform/shape
y_hat_src2/kernel/Initializer/random_uniform/min
y_hat_src2/kernel/Initializer/random_uniform/max
y_hat_src2/kernel/Initializer/random_uniform/RandomUniform
y_hat_src2/kernel/Initializer/random_uniform/sub
y_hat_src2/kernel/Initializer/random_uniform/mul
y_hat_src2/kernel/Initializer/random_uniform
y_hat_src2/kernel
y_hat_src2/kernel/Assign
y_hat_src2/kernel/read
y_hat_src2/bias/Initializer/zeros
y_hat_src2/bias
y_hat_src2/bias/Assign
y_hat_src2/bias/read
y_hat_src2/Tensordot/axes
y_hat_src2/Tensordot/free
y_hat_src2/Tensordot/Shape
y_hat_src2/Tensordot/GatherV2/axis
y_hat_src2/Tensordot/GatherV2
y_hat_src2/Tensordot/GatherV2_1/axis
y_hat_src2/Tensordot/GatherV2_1
y_hat_src2/Tensordot/Const
y_hat_src2/Tensordot/Prod
y_hat_src2/Tensordot/Const_1
y_hat_src2/Tensordot/Prod_1
y_hat_src2/Tensordot/concat/axis
y_hat_src2/Tensordot/concat
y_hat_src2/Tensordot/stack
y_hat_src2/Tensordot/transpose
y_hat_src2/Tensordot/Reshape
y_hat_src2/Tensordot/transpose_1/perm
y_hat_src2/Tensordot/transpose_1
y_hat_src2/Tensordot/Reshape_1/shape
y_hat_src2/Tensordot/Reshape_1
y_hat_src2/Tensordot/MatMul
y_hat_src2/Tensordot/Const_2
y_hat_src2/Tensordot/concat_1/axis
y_hat_src2/Tensordot/concat_1
y_hat_src2/Tensordot
y_hat_src2/BiasAdd
y_hat_src2/Relu
add
add_1/y
add_1
truediv
y_out1
add_2
add_3/y
add_3
truediv_1
y_out2
save/filename/input
save/filename
save/Const
save/SaveV2/tensor_names
save/SaveV2/shape_and_slices
save/SaveV2
save/control_dependency
save/RestoreV2/tensor_names
save/RestoreV2/shape_and_slices
save/RestoreV2
save/Assign
save/Assign_1
save/Assign_2
save/Assign_3
save/Assign_4
save/Assign_5
save/Assign_6
save/Assign_7
save/Assign_8
save/Assign_9
save/Assign_10
save/Assign_11
save/Assign_12
save/Assign_13
save/Assign_14
save/Assign_15
save/Assign_16
save/restore_all

 

W0906 16:33:20.548241 21784 deprecation.py:323] From D:/python/pycode/mss/clean_freeze2pb_m.py:23: convert_variables_to_constants (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.compat.v1.graph_util.convert_variables_to_constants`
W0906 16:33:20.570699 21784 deprecation.py:323] From D:\python\lib\site-packages\tensorflow\python\framework\graph_util_impl.py:270: extract_sub_graph (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.compat.v1.graph_util.extract_sub_graph`
I0906 16:33:20.629382 21784 graph_util_impl.py:311] Froze 16 variables.
I0906 16:33:20.677789 21784 graph_util_impl.py:364] Converted 16 variables to const ops.
W0906 16:33:20.696754 21784 deprecation_wrapper.py:119] From D:/python/pycode/mss/clean_freeze2pb_m.py:24: The name tf.train.write_graph is deprecated. Please use tf.io.write_graph instead.

W0906 16:33:20.834012 21784 deprecation_wrapper.py:119] From D:/python/pycode/mss/clean_freeze2pb_m.py:28: The name tf.logging.info is deprecated. Please use tf.compat.v1.logging.info instead.

I0906 16:33:20.885414 21784 clean_freeze2pb_m.py:28] Saved frozen graph to D:/python/pycode/mss/clean_ckpt_m/my_freeze_graph.pb

 

人生总是充满惊喜和意外，哈哈

真的好开心，比找到对象都开心，哈哈哈哈

 

另外有相关问题可以加入QQ群讨论，不设微信群

QQ群：868373192 

语音深度学习与信号处理群

 

小李飞刀李寻欢
关注
专栏目录
tensorflow定义自己的网络模型
12-20
tensorflow定义自己的网络模型，以MNIST数据集为例，实现网络的定义和训练
tensorflow自定义网络结构
Dean0731的博客
05-05 282
自定义网络层 自定义层需要继承tf.keras.layers.Layer类，重写init,build,call __init__,执行与输入无关的初始化 build，了解输入张量的形状，定义需要什么输入 call，进行正向计算 class MyDense(tf.keras.layers.Layer): def __init__(self,units): # units 神经...
tensorflow自定义网络层和结构
lzfshub
09-07 1910
1. 自定义网络层 实现自己的层的最佳方法是扩展tf.keras.layers.Layer类并实现： _init_()函数，在其中执行所有与输入无关的初始化 build()函数，获得输入张量的形状，并可以进行其余的初始化 call()函数，构建网络结构，进行前向传播 该自定以网络相当于：tf.keras.layers.Dense() 格式： class MyDense(tf.keras.layers.Layer): # 继承该类 def __init__(self, input_dim, outpu
【TensorFlow2.0】自定义网络层、自定义网络、内置预训练模型、测量工具
qq_36926037的博客
05-01 3402
一、自定义网络   网络层：   （1）常用网络层：keras提供了非常多的常用网络层类，用tf.keras.layers.xxx()实现，但深度学习可以使用的网络层远远不止这些。   （2）自定义网络层：需要创建自定义网络层类，这个类继承自tf.keras.layers.Layer基类；   网络（模型）：   （1）通常网络构建：利用tf.keras.Sequential()网络容器，将网路层...
tensorflow初学，生成一个自己的网络结构（代码）
11-20
tensorflow初学，生成一个自己的网络结构（代码），并进行训练
基本TensorFlow机制、操作、类定义和神经网络构建.zip
04-10
2. **构建模型**：使用tf.keras.layers构建网络结构，如输入层、隐藏层和输出层。 3. **定义损失函数**：选择合适的损失函数，如交叉熵或均方误差。 4. **选择优化器**：选择优化算法，如Adam或SGD，用于最小化损失...
TensorFlow实现基于遗传算法的神经网络结构搜索技术
最新发布
03-02
4. **目标函数**：在基于遗传算法的神经网络结构搜索中，需要定义一个适应度函数或目标函数，用于评估每个神经网络的性能。目标函数可以是训练集上的准确率、验证集上的损失值等，根据具体任务和要求来确定。 ........
BP神经网络实例及代码分析(python+tensorflow),bp神经网络例子,Python
09-10
**五、实例应用**：在具体的问题中，如分类或回归任务，我们可以根据BP神经网络的框架，调整网络结构和参数，以适应特定问题的需求。 **六、常见问题与解决策略** 1. **过拟合**：通过正则化或早停法来防止过拟合...
TensorFlow搭建神经网络最佳实践
09-20
在本文中，我们将深入探讨如何使用TensorFlow框架搭建神经网络的最佳实践。TensorFlow是一个强大的开源库，用于构建和训练各种机器学习模型，包括神经网络。在这个实践中，我们以经典的MNIST手写数字识别数据集为例...
tensorflow实现残差网络方式(mnist数据集)
09-16
在本文中，我们将深入探讨如何使用TensorFlow框架实现残差网络（ResNet）来处理MNIST数据...尽管这只是一个基础的实现，但它展示了残差网络的基本原理，你可以根据实际需求增加层数或调整网络结构以适应更复杂的任务。
tensorflow 重置/清除计算图的实现
09-18
今天小编就为大家分享一篇tensorflow 重置/清除计算图的实现，具有很好的参考价值，希望对大家有所帮助。一起跟随小编过来看看吧
Tensorflow2.0 自定义网络
廷益_飞鸟的博客
10-19 1087
自定义网络 keras.Sequential 容器 keras.layers.Layer keras.Model Keras.Sequential 容器 网络层的搭建 model = keras.Sequential([keras.layers.Dense(256, activation="relu"), keras.layers.Dense(128, activation="relu"), kera
TensorFlow2.0:自定义层与自定义网络
meijie2018_1的博客
09-11 3245
自定义层函数需要继承layers.Layer，自定义网络需要继承keras.Model。 其内部需要定义两个函数： 1、__init__初始化函数，内部需要定义构造形式； 2、call函数，内部需要定义计算形式及返回值。 #self def layer class MyDense(layers.Layer):#inherit layers.Layer def __init__(self,i...
TensorFlow中的图（自定义图）
大脸猫的博客
01-21 993
上一节（TensorFlow2.0之计算过程与名字来历）中我们反复提到了图，但是代码中并没有看到图的定义，也没看到任何跟图有关的代码。其实，在TensorFlow中，TensorFlow会定义默认图。用户可以自己显式定义图，并将自定义图作为默认图。 TensorFlow的图中包含tf.Operation对象集合，一个tf.Operation对象表示一个计算单元，如加、减、乘、除等计算都是tf.O...
tensorflow自定义结构初步
qq_42778110的博客
01-27 307
0.摘要 对输入值x，标签y，模型参数c，我们希望习得一个最优的Model-c满足y=Model(x,c)。机器学习的任务就是要确定这个Model中的参数。但是因为Model的形状(如层数，节点个数等)是离散的，处理起来相对麻烦，所以我们在这里只讨论可导的参数c(tensorflow的一个优势就在于可以自动求导)。我们希望对于一个确定确定的模型Model，我们要找到这样的c满足c=argmin(l...
tensorflow的简单网络结构程序
zjguilai的博客
05-15 428
1）简单优化器的构建代码： import tensorflow as tf import numpy as np #create data #生成100个一维随机数列 x_data = np.random.rand(100).astype(np.float32) #这里表示我们要预测的Weight和biases y_data = x_data*0.1 + 0.3 **##create tenso...
tensorflow keras API定义神经网络结构
hangsyt108的博客
06-21 479
Keras有三种API，分别是Sequential,functional,和SubclassingAPI，只介绍前两种。 一、Sequential API # Define the inference model model = tf.keras.Sequential() model.add(tf.keras.layers.Dense(256, activation=tf.nn.tanh))...
Tensorflow 网络基本结构
zzj的博客
12-11 1093
Tensorflow 网络基本结构
tensorflow网络构建结构入门
霜叶的博客
09-26 1098
1、tensorflow中的神经网络表示 数据：张量表示数据 网络：计算图搭建网络；会话执行计算图，优化线上权重(参数)，得到模型 2、张量 张量通熟地讲就是多维数组 （列表） 3、数据类型 常用的数据类型有tf.float32 ,tf.int32; 4、计算图 计算图是用来搭建神经网络的计算过程，只搭建，不运算 5、Session Session表示会话，用来执行计算图中的节点运算； 6、参数 7、构建神经网络 构建神经网络过程： 数据准备-->搭
tensorflow定义fcn网络结构
04-28
下面是一个简单的FCN网络结构的TensorFlow实现示例： ```python import tensorflow as tf def fcn(images, num_classes): # 编码器部分 conv1 = tf.layers.conv2d(images, filters=64, kernel_size=3, strides=1...
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包

打赏作者

小李飞刀李寻欢

您的欣赏将是我奋斗路上的动力！

¥1 ¥2 ¥4 ¥6 ¥10 ¥20
扫码支付：¥1
获取中
扫码支付

您的余额不足，请更换扫码支付或充值

打赏作者

实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值