Model: "encoder_r1" __________________________________________________________________________________________________ Layer (type) Output Shape Param # Connected to ================================================================================================== input_1 (InputLayer) [(None, 4096, 3)] 0 __________________________________________________________________________________________________ shared_conv_0 (Conv1D) (None, 4096, 96) 18528 input_1[0][0] __________________________________________________________________________________________________ shared_conv_0_batchnorm (BatchN (None, 4096, 96) 384 shared_conv_0[0][0] __________________________________________________________________________________________________ leaky_re_lu (LeakyReLU) multiple 0 shared_conv_0_batchnorm[0][0] shared_conv_2_batchnorm[0][0] shared_conv_4_batchnorm[0][0] shared_conv_6_batchnorm[0][0] shared_conv_8_batchnorm[0][0] shared_conv_10_batchnorm[0][0] r1_conv2d_1_batchnorm[0][0] batch_normalization[0][0] batch_normalization_1[0][0] add[0][0] batch_normalization_3[0][0] batch_normalization_4[0][0] add_1[0][0] batch_normalization_6[0][0] batch_normalization_7[0][0] add_2[0][0] r1_dense_9_batchnorm[0][0] __________________________________________________________________________________________________ max_pooling1d (MaxPooling1D) (None, 2048, 96) 0 leaky_re_lu[0][0] __________________________________________________________________________________________________ shared_conv_2 (Conv1D) (None, 2048, 64) 393280 max_pooling1d[0][0] __________________________________________________________________________________________________ shared_conv_2_batchnorm (BatchN (None, 2048, 64) 256 shared_conv_2[0][0] __________________________________________________________________________________________________ max_pooling1d_1 (MaxPooling1D) (None, 1024, 64) 0 leaky_re_lu[1][0] __________________________________________________________________________________________________ shared_conv_4 (Conv1D) (None, 1024, 64) 131136 max_pooling1d_1[0][0] __________________________________________________________________________________________________ shared_conv_4_batchnorm (BatchN (None, 1024, 64) 256 shared_conv_4[0][0] __________________________________________________________________________________________________ max_pooling1d_2 (MaxPooling1D) (None, 512, 64) 0 leaky_re_lu[2][0] __________________________________________________________________________________________________ shared_conv_6 (Conv1D) (None, 512, 64) 131136 max_pooling1d_2[0][0] __________________________________________________________________________________________________ shared_conv_6_batchnorm (BatchN (None, 512, 64) 256 shared_conv_6[0][0] __________________________________________________________________________________________________ max_pooling1d_3 (MaxPooling1D) (None, 256, 64) 0 leaky_re_lu[3][0] __________________________________________________________________________________________________ shared_conv_8 (Conv1D) (None, 256, 64) 65600 max_pooling1d_3[0][0] __________________________________________________________________________________________________ shared_conv_8_batchnorm (BatchN (None, 256, 64) 256 shared_conv_8[0][0] __________________________________________________________________________________________________ max_pooling1d_4 (MaxPooling1D) (None, 128, 64) 0 leaky_re_lu[4][0] __________________________________________________________________________________________________ shared_conv_10 (Conv1D) (None, 128, 64) 65600 max_pooling1d_4[0][0] __________________________________________________________________________________________________ shared_conv_10_batchnorm (Batch (None, 128, 64) 256 shared_conv_10[0][0] __________________________________________________________________________________________________ max_pooling1d_5 (MaxPooling1D) (None, 64, 64) 0 leaky_re_lu[5][0] __________________________________________________________________________________________________ reshape (Reshape) (None, 64, 64, 1) 0 max_pooling1d_5[0][0] __________________________________________________________________________________________________ r1_conv2d_1 (Conv2D) (None, 64, 64, 32) 832 reshape[0][0] __________________________________________________________________________________________________ r1_conv2d_1_batchnorm (BatchNor (None, 64, 64, 32) 128 r1_conv2d_1[0][0] __________________________________________________________________________________________________ conv2d (Conv2D) (None, 64, 64, 32) 1056 leaky_re_lu[6][0] __________________________________________________________________________________________________ batch_normalization (BatchNorma (None, 64, 64, 32) 128 conv2d[0][0] __________________________________________________________________________________________________ conv2d_1 (Conv2D) (None, 64, 64, 32) 25632 leaky_re_lu[7][0] __________________________________________________________________________________________________ batch_normalization_1 (BatchNor (None, 64, 64, 32) 128 conv2d_1[0][0] __________________________________________________________________________________________________ conv2d_2 (Conv2D) (None, 64, 64, 32) 1056 leaky_re_lu[8][0] __________________________________________________________________________________________________ batch_normalization_2 (BatchNor (None, 64, 64, 32) 128 conv2d_2[0][0] __________________________________________________________________________________________________ add (Add) (None, 64, 64, 32) 0 batch_normalization_2[0][0] leaky_re_lu[6][0] __________________________________________________________________________________________________ max_pooling2d (MaxPooling2D) (None, 32, 32, 32) 0 leaky_re_lu[9][0] __________________________________________________________________________________________________ conv2d_3 (Conv2D) (None, 32, 32, 32) 1056 max_pooling2d[0][0] __________________________________________________________________________________________________ batch_normalization_3 (BatchNor (None, 32, 32, 32) 128 conv2d_3[0][0] __________________________________________________________________________________________________ conv2d_4 (Conv2D) (None, 32, 32, 32) 9248 leaky_re_lu[10][0] __________________________________________________________________________________________________ batch_normalization_4 (BatchNor (None, 32, 32, 32) 128 conv2d_4[0][0] __________________________________________________________________________________________________ conv2d_5 (Conv2D) (None, 32, 32, 32) 1056 leaky_re_lu[11][0] __________________________________________________________________________________________________ batch_normalization_5 (BatchNor (None, 32, 32, 32) 128 conv2d_5[0][0] __________________________________________________________________________________________________ add_1 (Add) (None, 32, 32, 32) 0 batch_normalization_5[0][0] max_pooling2d[0][0] __________________________________________________________________________________________________ max_pooling2d_1 (MaxPooling2D) (None, 16, 16, 32) 0 leaky_re_lu[12][0] __________________________________________________________________________________________________ conv2d_6 (Conv2D) (None, 16, 16, 32) 1056 max_pooling2d_1[0][0] __________________________________________________________________________________________________ batch_normalization_6 (BatchNor (None, 16, 16, 32) 128 conv2d_6[0][0] __________________________________________________________________________________________________ conv2d_7 (Conv2D) (None, 16, 16, 32) 9248 leaky_re_lu[13][0] __________________________________________________________________________________________________ batch_normalization_7 (BatchNor (None, 16, 16, 32) 128 conv2d_7[0][0] __________________________________________________________________________________________________ conv2d_8 (Conv2D) (None, 16, 16, 32) 1056 leaky_re_lu[14][0] __________________________________________________________________________________________________ batch_normalization_8 (BatchNor (None, 16, 16, 32) 128 conv2d_8[0][0] __________________________________________________________________________________________________ add_2 (Add) (None, 16, 16, 32) 0 batch_normalization_8[0][0] max_pooling2d_1[0][0] __________________________________________________________________________________________________ max_pooling2d_2 (MaxPooling2D) (None, 8, 8, 32) 0 leaky_re_lu[15][0] __________________________________________________________________________________________________ flatten (Flatten) (None, 2048) 0 max_pooling2d_2[0][0] __________________________________________________________________________________________________ r1_dense_9 (Dense) (None, 1024) 2098176 flatten[0][0] __________________________________________________________________________________________________ r1_dense_9_batchnorm (BatchNorm (None, 1024) 4096 r1_dense_9[0][0] __________________________________________________________________________________________________ r1_mean_dense (Dense) (None, 6) 6150 leaky_re_lu[16][0] __________________________________________________________________________________________________ r1_logvar_dense (Dense) (None, 6) 6150 leaky_re_lu[16][0] __________________________________________________________________________________________________ r1_modes_dense (Dense) (None, 3) 3075 leaky_re_lu[16][0] __________________________________________________________________________________________________ concatenate (Concatenate) (None, 15) 0 r1_mean_dense[0][0] r1_logvar_dense[0][0] r1_modes_dense[0][0] ================================================================================================== Total params: 2,977,167 Trainable params: 2,973,647 Non-trainable params: 3,520 __________________________________________________________________________________________________ Model: "encoder_q" __________________________________________________________________________________________________ Layer (type) Output Shape Param # Connected to ================================================================================================== input_1 (InputLayer) [(None, 4096, 3)] 0 __________________________________________________________________________________________________ shared_conv_0 (Conv1D) (None, 4096, 96) 18528 input_1[0][0] __________________________________________________________________________________________________ shared_conv_0_batchnorm (BatchN (None, 4096, 96) 384 shared_conv_0[0][0] __________________________________________________________________________________________________ leaky_re_lu (LeakyReLU) multiple 0 shared_conv_0_batchnorm[0][0] shared_conv_2_batchnorm[0][0] shared_conv_4_batchnorm[0][0] shared_conv_6_batchnorm[0][0] shared_conv_8_batchnorm[0][0] shared_conv_10_batchnorm[0][0] batch_normalization_9[0][0] q_conv2d_1_batchnorm[0][0] batch_normalization_10[0][0] batch_normalization_11[0][0] add_3[0][0] batch_normalization_13[0][0] batch_normalization_14[0][0] add_4[0][0] batch_normalization_16[0][0] batch_normalization_17[0][0] add_5[0][0] q_dense_9_batchnorm[0][0] __________________________________________________________________________________________________ max_pooling1d (MaxPooling1D) (None, 2048, 96) 0 leaky_re_lu[0][0] __________________________________________________________________________________________________ shared_conv_2 (Conv1D) (None, 2048, 64) 393280 max_pooling1d[0][0] __________________________________________________________________________________________________ shared_conv_2_batchnorm (BatchN (None, 2048, 64) 256 shared_conv_2[0][0] __________________________________________________________________________________________________ max_pooling1d_1 (MaxPooling1D) (None, 1024, 64) 0 leaky_re_lu[1][0] __________________________________________________________________________________________________ shared_conv_4 (Conv1D) (None, 1024, 64) 131136 max_pooling1d_1[0][0] __________________________________________________________________________________________________ shared_conv_4_batchnorm (BatchN (None, 1024, 64) 256 shared_conv_4[0][0] __________________________________________________________________________________________________ max_pooling1d_2 (MaxPooling1D) (None, 512, 64) 0 leaky_re_lu[2][0] __________________________________________________________________________________________________ shared_conv_6 (Conv1D) (None, 512, 64) 131136 max_pooling1d_2[0][0] __________________________________________________________________________________________________ shared_conv_6_batchnorm (BatchN (None, 512, 64) 256 shared_conv_6[0][0] __________________________________________________________________________________________________ max_pooling1d_3 (MaxPooling1D) (None, 256, 64) 0 leaky_re_lu[3][0] __________________________________________________________________________________________________ shared_conv_8 (Conv1D) (None, 256, 64) 65600 max_pooling1d_3[0][0] __________________________________________________________________________________________________ shared_conv_8_batchnorm (BatchN (None, 256, 64) 256 shared_conv_8[0][0] __________________________________________________________________________________________________ max_pooling1d_4 (MaxPooling1D) (None, 128, 64) 0 leaky_re_lu[4][0] __________________________________________________________________________________________________ input_2 (InputLayer) [(None, 2)] 0 __________________________________________________________________________________________________ shared_conv_10 (Conv1D) (None, 128, 64) 65600 max_pooling1d_4[0][0] __________________________________________________________________________________________________ flatten_1 (Flatten) (None, 2) 0 input_2[0][0] __________________________________________________________________________________________________ shared_conv_10_batchnorm (Batch (None, 128, 64) 256 shared_conv_10[0][0] __________________________________________________________________________________________________ q_inx_dense (Dense) (None, 4096) 12288 flatten_1[0][0] __________________________________________________________________________________________________ tf.reshape (TFOpLambda) (None, 64, 64, 1) 0 q_inx_dense[0][0] __________________________________________________________________________________________________ max_pooling1d_5 (MaxPooling1D) (None, 64, 64) 0 leaky_re_lu[5][0] __________________________________________________________________________________________________ batch_normalization_9 (BatchNor (None, 64, 64, 1) 4 tf.reshape[0][0] __________________________________________________________________________________________________ tf.expand_dims (TFOpLambda) (None, 64, 64, 1) 0 max_pooling1d_5[0][0] __________________________________________________________________________________________________ concatenate_1 (Concatenate) (None, 64, 64, 2) 0 tf.expand_dims[0][0] leaky_re_lu[17][0] __________________________________________________________________________________________________ reshape_1 (Reshape) (None, 64, 64, 2) 0 concatenate_1[0][0] __________________________________________________________________________________________________ q_conv2d_1 (Conv2D) (None, 64, 64, 32) 1632 reshape_1[0][0] __________________________________________________________________________________________________ q_conv2d_1_batchnorm (BatchNorm (None, 64, 64, 32) 128 q_conv2d_1[0][0] __________________________________________________________________________________________________ conv2d_9 (Conv2D) (None, 64, 64, 32) 1056 leaky_re_lu[18][0] __________________________________________________________________________________________________ batch_normalization_10 (BatchNo (None, 64, 64, 32) 128 conv2d_9[0][0] __________________________________________________________________________________________________ conv2d_10 (Conv2D) (None, 64, 64, 32) 25632 leaky_re_lu[19][0] __________________________________________________________________________________________________ batch_normalization_11 (BatchNo (None, 64, 64, 32) 128 conv2d_10[0][0] __________________________________________________________________________________________________ conv2d_11 (Conv2D) (None, 64, 64, 32) 1056 leaky_re_lu[20][0] __________________________________________________________________________________________________ batch_normalization_12 (BatchNo (None, 64, 64, 32) 128 conv2d_11[0][0] __________________________________________________________________________________________________ add_3 (Add) (None, 64, 64, 32) 0 batch_normalization_12[0][0] leaky_re_lu[18][0] __________________________________________________________________________________________________ max_pooling2d_3 (MaxPooling2D) (None, 32, 32, 32) 0 leaky_re_lu[21][0] __________________________________________________________________________________________________ conv2d_12 (Conv2D) (None, 32, 32, 32) 1056 max_pooling2d_3[0][0] __________________________________________________________________________________________________ batch_normalization_13 (BatchNo (None, 32, 32, 32) 128 conv2d_12[0][0] __________________________________________________________________________________________________ conv2d_13 (Conv2D) (None, 32, 32, 32) 9248 leaky_re_lu[22][0] __________________________________________________________________________________________________ batch_normalization_14 (BatchNo (None, 32, 32, 32) 128 conv2d_13[0][0] __________________________________________________________________________________________________ conv2d_14 (Conv2D) (None, 32, 32, 32) 1056 leaky_re_lu[23][0] __________________________________________________________________________________________________ batch_normalization_15 (BatchNo (None, 32, 32, 32) 128 conv2d_14[0][0] __________________________________________________________________________________________________ add_4 (Add) (None, 32, 32, 32) 0 batch_normalization_15[0][0] max_pooling2d_3[0][0] __________________________________________________________________________________________________ max_pooling2d_4 (MaxPooling2D) (None, 16, 16, 32) 0 leaky_re_lu[24][0] __________________________________________________________________________________________________ conv2d_15 (Conv2D) (None, 16, 16, 32) 1056 max_pooling2d_4[0][0] __________________________________________________________________________________________________ batch_normalization_16 (BatchNo (None, 16, 16, 32) 128 conv2d_15[0][0] __________________________________________________________________________________________________ conv2d_16 (Conv2D) (None, 16, 16, 32) 9248 leaky_re_lu[25][0] __________________________________________________________________________________________________ batch_normalization_17 (BatchNo (None, 16, 16, 32) 128 conv2d_16[0][0] __________________________________________________________________________________________________ conv2d_17 (Conv2D) (None, 16, 16, 32) 1056 leaky_re_lu[26][0] __________________________________________________________________________________________________ batch_normalization_18 (BatchNo (None, 16, 16, 32) 128 conv2d_17[0][0] __________________________________________________________________________________________________ add_5 (Add) (None, 16, 16, 32) 0 batch_normalization_18[0][0] max_pooling2d_4[0][0] __________________________________________________________________________________________________ max_pooling2d_5 (MaxPooling2D) (None, 8, 8, 32) 0 leaky_re_lu[27][0] __________________________________________________________________________________________________ flatten_2 (Flatten) (None, 2048) 0 max_pooling2d_5[0][0] __________________________________________________________________________________________________ q_dense_9 (Dense) (None, 1024) 2098176 flatten_2[0][0] __________________________________________________________________________________________________ q_dense_9_batchnorm (BatchNorma (None, 1024) 4096 q_dense_9[0][0] __________________________________________________________________________________________________ q_mean_dense (Dense) (None, 2) 2050 leaky_re_lu[28][0] __________________________________________________________________________________________________ q_logvar_dense (Dense) (None, 2) 2050 leaky_re_lu[28][0] __________________________________________________________________________________________________ concatenate_2 (Concatenate) (None, 4) 0 q_mean_dense[0][0] q_logvar_dense[0][0] ================================================================================================== Total params: 2,978,984 Trainable params: 2,975,462 Non-trainable params: 3,522 __________________________________________________________________________________________________ Model: "decoder_r2" __________________________________________________________________________________________________ Layer (type) Output Shape Param # Connected to ================================================================================================== input_1 (InputLayer) [(None, 4096, 3)] 0 __________________________________________________________________________________________________ shared_conv_0 (Conv1D) (None, 4096, 96) 18528 input_1[0][0] __________________________________________________________________________________________________ shared_conv_0_batchnorm (BatchN (None, 4096, 96) 384 shared_conv_0[0][0] __________________________________________________________________________________________________ leaky_re_lu (LeakyReLU) multiple 0 shared_conv_0_batchnorm[0][0] shared_conv_2_batchnorm[0][0] shared_conv_4_batchnorm[0][0] shared_conv_6_batchnorm[0][0] shared_conv_8_batchnorm[0][0] shared_conv_10_batchnorm[0][0] batch_normalization_19[0][0] r2_conv2d_1_batchnorm[0][0] batch_normalization_20[0][0] batch_normalization_21[0][0] add_6[0][0] batch_normalization_23[0][0] batch_normalization_24[0][0] add_7[0][0] batch_normalization_26[0][0] batch_normalization_27[0][0] add_8[0][0] r2_dense_9_batchnorm[0][0] __________________________________________________________________________________________________ max_pooling1d (MaxPooling1D) (None, 2048, 96) 0 leaky_re_lu[0][0] __________________________________________________________________________________________________ shared_conv_2 (Conv1D) (None, 2048, 64) 393280 max_pooling1d[0][0] __________________________________________________________________________________________________ shared_conv_2_batchnorm (BatchN (None, 2048, 64) 256 shared_conv_2[0][0] __________________________________________________________________________________________________ max_pooling1d_1 (MaxPooling1D) (None, 1024, 64) 0 leaky_re_lu[1][0] __________________________________________________________________________________________________ shared_conv_4 (Conv1D) (None, 1024, 64) 131136 max_pooling1d_1[0][0] __________________________________________________________________________________________________ shared_conv_4_batchnorm (BatchN (None, 1024, 64) 256 shared_conv_4[0][0] __________________________________________________________________________________________________ max_pooling1d_2 (MaxPooling1D) (None, 512, 64) 0 leaky_re_lu[2][0] __________________________________________________________________________________________________ shared_conv_6 (Conv1D) (None, 512, 64) 131136 max_pooling1d_2[0][0] __________________________________________________________________________________________________ shared_conv_6_batchnorm (BatchN (None, 512, 64) 256 shared_conv_6[0][0] __________________________________________________________________________________________________ max_pooling1d_3 (MaxPooling1D) (None, 256, 64) 0 leaky_re_lu[3][0] __________________________________________________________________________________________________ shared_conv_8 (Conv1D) (None, 256, 64) 65600 max_pooling1d_3[0][0] __________________________________________________________________________________________________ shared_conv_8_batchnorm (BatchN (None, 256, 64) 256 shared_conv_8[0][0] __________________________________________________________________________________________________ max_pooling1d_4 (MaxPooling1D) (None, 128, 64) 0 leaky_re_lu[4][0] __________________________________________________________________________________________________ input_3 (InputLayer) [(None, 2)] 0 __________________________________________________________________________________________________ shared_conv_10 (Conv1D) (None, 128, 64) 65600 max_pooling1d_4[0][0] __________________________________________________________________________________________________ flatten_3 (Flatten) (None, 2) 0 input_3[0][0] __________________________________________________________________________________________________ shared_conv_10_batchnorm (Batch (None, 128, 64) 256 shared_conv_10[0][0] __________________________________________________________________________________________________ r2_inz_dense (Dense) (None, 4096) 12288 flatten_3[0][0] __________________________________________________________________________________________________ tf.reshape_1 (TFOpLambda) (None, 64, 64, 1) 0 r2_inz_dense[0][0] __________________________________________________________________________________________________ max_pooling1d_5 (MaxPooling1D) (None, 64, 64) 0 leaky_re_lu[5][0] __________________________________________________________________________________________________ batch_normalization_19 (BatchNo (None, 64, 64, 1) 4 tf.reshape_1[0][0] __________________________________________________________________________________________________ tf.expand_dims_1 (TFOpLambda) (None, 64, 64, 1) 0 max_pooling1d_5[0][0] __________________________________________________________________________________________________ concatenate_3 (Concatenate) (None, 64, 64, 2) 0 tf.expand_dims_1[0][0] leaky_re_lu[29][0] __________________________________________________________________________________________________ reshape_2 (Reshape) (None, 64, 64, 2) 0 concatenate_3[0][0] __________________________________________________________________________________________________ r2_conv2d_1 (Conv2D) (None, 64, 64, 32) 1632 reshape_2[0][0] __________________________________________________________________________________________________ r2_conv2d_1_batchnorm (BatchNor (None, 64, 64, 32) 128 r2_conv2d_1[0][0] __________________________________________________________________________________________________ conv2d_18 (Conv2D) (None, 64, 64, 32) 1056 leaky_re_lu[30][0] __________________________________________________________________________________________________ batch_normalization_20 (BatchNo (None, 64, 64, 32) 128 conv2d_18[0][0] __________________________________________________________________________________________________ conv2d_19 (Conv2D) (None, 64, 64, 32) 25632 leaky_re_lu[31][0] __________________________________________________________________________________________________ batch_normalization_21 (BatchNo (None, 64, 64, 32) 128 conv2d_19[0][0] __________________________________________________________________________________________________ conv2d_20 (Conv2D) (None, 64, 64, 32) 1056 leaky_re_lu[32][0] __________________________________________________________________________________________________ batch_normalization_22 (BatchNo (None, 64, 64, 32) 128 conv2d_20[0][0] __________________________________________________________________________________________________ add_6 (Add) (None, 64, 64, 32) 0 batch_normalization_22[0][0] leaky_re_lu[30][0] __________________________________________________________________________________________________ max_pooling2d_6 (MaxPooling2D) (None, 32, 32, 32) 0 leaky_re_lu[33][0] __________________________________________________________________________________________________ conv2d_21 (Conv2D) (None, 32, 32, 32) 1056 max_pooling2d_6[0][0] __________________________________________________________________________________________________ batch_normalization_23 (BatchNo (None, 32, 32, 32) 128 conv2d_21[0][0] __________________________________________________________________________________________________ conv2d_22 (Conv2D) (None, 32, 32, 32) 9248 leaky_re_lu[34][0] __________________________________________________________________________________________________ batch_normalization_24 (BatchNo (None, 32, 32, 32) 128 conv2d_22[0][0] __________________________________________________________________________________________________ conv2d_23 (Conv2D) (None, 32, 32, 32) 1056 leaky_re_lu[35][0] __________________________________________________________________________________________________ batch_normalization_25 (BatchNo (None, 32, 32, 32) 128 conv2d_23[0][0] __________________________________________________________________________________________________ add_7 (Add) (None, 32, 32, 32) 0 batch_normalization_25[0][0] max_pooling2d_6[0][0] __________________________________________________________________________________________________ max_pooling2d_7 (MaxPooling2D) (None, 16, 16, 32) 0 leaky_re_lu[36][0] __________________________________________________________________________________________________ conv2d_24 (Conv2D) (None, 16, 16, 32) 1056 max_pooling2d_7[0][0] __________________________________________________________________________________________________ batch_normalization_26 (BatchNo (None, 16, 16, 32) 128 conv2d_24[0][0] __________________________________________________________________________________________________ conv2d_25 (Conv2D) (None, 16, 16, 32) 9248 leaky_re_lu[37][0] __________________________________________________________________________________________________ batch_normalization_27 (BatchNo (None, 16, 16, 32) 128 conv2d_25[0][0] __________________________________________________________________________________________________ conv2d_26 (Conv2D) (None, 16, 16, 32) 1056 leaky_re_lu[38][0] __________________________________________________________________________________________________ batch_normalization_28 (BatchNo (None, 16, 16, 32) 128 conv2d_26[0][0] __________________________________________________________________________________________________ add_8 (Add) (None, 16, 16, 32) 0 batch_normalization_28[0][0] max_pooling2d_7[0][0] __________________________________________________________________________________________________ max_pooling2d_8 (MaxPooling2D) (None, 8, 8, 32) 0 leaky_re_lu[39][0] __________________________________________________________________________________________________ flatten_4 (Flatten) (None, 2048) 0 max_pooling2d_8[0][0] __________________________________________________________________________________________________ r2_dense_9 (Dense) (None, 1024) 2098176 flatten_4[0][0] __________________________________________________________________________________________________ r2_dense_9_batchnorm (BatchNorm (None, 1024) 4096 r2_dense_9[0][0] __________________________________________________________________________________________________ JointVonMisesFisher_mean (Dense (None, 3) 3075 leaky_re_lu[40][0] __________________________________________________________________________________________________ JointVonMisesFisher_logvar (Den (None, 1) 1025 leaky_re_lu[40][0] __________________________________________________________________________________________________ concatenate_4 (Concatenate) (None, 4) 0 JointVonMisesFisher_mean[0][0] JointVonMisesFisher_logvar[0][0] ================================================================================================== Total params: 2,978,984 Trainable params: 2,975,462 Non-trainable params: 3,522 __________________________________________________________________________________________________