Comments (8)
Output:
Extracting 60000 images of 28x28 from train-images-idx3-ubyte.gz
Extracting 60000 labels from train-labels-idx1-ubyte.gz
Extracting 10000 images of 28x28 from t10k-images-idx3-ubyte.gz
Extracting 10000 labels from t10k-labels-idx1-ubyte.gz
21:49:39.788 [main] DEBUG api.keras.Sequential - Conv2D(filters=32, kernelSize=[5, 5], strides=[1, 1, 1, 1], dilations=[1, 1, 1, 1], activation=Relu, kernelInitializer=HeNormal(seed=12) VarianceScaling(scale=2.0, mode=FAN_IN, distribution=TRUNCATED_NORMAL, seed=12), biasInitializer=api.keras.initializers.Zeros@be35cd9, kernelShape=[5, 5, 1, 32], padding=SAME); outputShape: [-1, 28, 28, 32]
21:49:39.789 [main] DEBUG api.keras.Sequential - MaxPool2D(poolSize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding=SAME); outputShape: [-1, 14, 14, 32]
21:49:39.790 [main] DEBUG api.keras.Sequential - Conv2D(filters=64, kernelSize=[5, 5], strides=[1, 1, 1, 1], dilations=[1, 1, 1, 1], activation=Relu, kernelInitializer=HeNormal(seed=12) VarianceScaling(scale=2.0, mode=FAN_IN, distribution=TRUNCATED_NORMAL, seed=12), biasInitializer=api.keras.initializers.Zeros@1b6e1eff, kernelShape=[5, 5, 32, 64], padding=SAME); outputShape: [-1, 14, 14, 64]
21:49:39.790 [main] DEBUG api.keras.Sequential - MaxPool2D(poolSize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding=SAME); outputShape: [-1, 7, 7, 64]
21:49:39.791 [main] DEBUG api.keras.Sequential - Flatten; outputShape: [3136]
21:49:39.792 [main] DEBUG api.keras.Sequential - Dense(outputSize=512, activation=Relu, kernelInitializer=HeNormal(seed=12) VarianceScaling(scale=2.0, mode=FAN_IN, distribution=TRUNCATED_NORMAL, seed=12), biasInitializer=Constant(constantValue=0.1), kernelShape=[3136, 512], biasShape=[512]); outputShape: [512]
21:49:39.793 [main] DEBUG api.keras.Sequential - Dense(outputSize=10, activation=Linear, kernelInitializer=HeNormal(seed=12) VarianceScaling(scale=2.0, mode=FAN_IN, distribution=TRUNCATED_NORMAL, seed=12), biasInitializer=Constant(constantValue=0.1), kernelShape=[512, 10], biasShape=[10]); outputShape: [10]
Name: default_data_placeholder; Type: Placeholder; Out #tensors: 1
Name: conv2d_1_conv2d_kernel; Type: VariableV2; Out #tensors: 1
Name: conv2d_1_conv2d_bias; Type: VariableV2; Out #tensors: 1
Name: Const; Type: Const; Out #tensors: 1
Name: Const_1; Type: Const; Out #tensors: 1
Name: StatelessTruncatedNormal; Type: StatelessTruncatedNormal; Out #tensors: 1
Name: Const_2; Type: Const; Out #tensors: 1
Name: Cast; Type: Cast; Out #tensors: 1
Name: Init_conv2d_1_conv2d_kernel; Type: Mul; Out #tensors: 1
Name: Assign_conv2d_1_conv2d_kernel; Type: Assign; Out #tensors: 1
Name: Const_3; Type: Const; Out #tensors: 1
Name: Init_conv2d_1_conv2d_bias/Zero; Type: Const; Out #tensors: 1
Name: Init_conv2d_1_conv2d_bias/Fill; Type: Fill; Out #tensors: 1
Name: Assign_conv2d_1_conv2d_bias; Type: Assign; Out #tensors: 1
Name: conv2d_3_conv2d_kernel; Type: VariableV2; Out #tensors: 1
Name: conv2d_3_conv2d_bias; Type: VariableV2; Out #tensors: 1
Name: Const_4; Type: Const; Out #tensors: 1
Name: Const_5; Type: Const; Out #tensors: 1
Name: StatelessTruncatedNormal_1; Type: StatelessTruncatedNormal; Out #tensors: 1
Name: Const_6; Type: Const; Out #tensors: 1
Name: Cast_1; Type: Cast; Out #tensors: 1
Name: Init_conv2d_3_conv2d_kernel; Type: Mul; Out #tensors: 1
Name: Assign_conv2d_3_conv2d_kernel; Type: Assign; Out #tensors: 1
Name: Const_7; Type: Const; Out #tensors: 1
Name: Init_conv2d_3_conv2d_bias/Zero; Type: Const; Out #tensors: 1
Name: Init_conv2d_3_conv2d_bias/Fill; Type: Fill; Out #tensors: 1
Name: Assign_conv2d_3_conv2d_bias; Type: Assign; Out #tensors: 1
Name: Const_8; Type: Const; Out #tensors: 1
Name: dense_6_dense_kernel; Type: VariableV2; Out #tensors: 1
Name: dense_6_dense_bias; Type: VariableV2; Out #tensors: 1
Name: Const_9; Type: Const; Out #tensors: 1
Name: Const_10; Type: Const; Out #tensors: 1
Name: StatelessTruncatedNormal_2; Type: StatelessTruncatedNormal; Out #tensors: 1
Name: Const_11; Type: Const; Out #tensors: 1
Name: Cast_2; Type: Cast; Out #tensors: 1
Name: Init_dense_6_dense_kernel; Type: Mul; Out #tensors: 1
Name: Assign_dense_6_dense_kernel; Type: Assign; Out #tensors: 1
Name: Const_12; Type: Const; Out #tensors: 1
Name: Const_13; Type: Const; Out #tensors: 1
Name: Init_dense_6_dense_bias; Type: Fill; Out #tensors: 1
Name: Assign_dense_6_dense_bias; Type: Assign; Out #tensors: 1
Name: dense_7_dense_kernel; Type: VariableV2; Out #tensors: 1
Name: dense_7_dense_bias; Type: VariableV2; Out #tensors: 1
Name: Const_14; Type: Const; Out #tensors: 1
Name: Const_15; Type: Const; Out #tensors: 1
Name: StatelessTruncatedNormal_3; Type: StatelessTruncatedNormal; Out #tensors: 1
Name: Const_16; Type: Const; Out #tensors: 1
Name: Cast_3; Type: Cast; Out #tensors: 1
Name: Init_dense_7_dense_kernel; Type: Mul; Out #tensors: 1
Name: Assign_dense_7_dense_kernel; Type: Assign; Out #tensors: 1
Name: Const_17; Type: Const; Out #tensors: 1
Name: Const_18; Type: Const; Out #tensors: 1
Name: Init_dense_7_dense_bias; Type: Fill; Out #tensors: 1
Name: Assign_dense_7_dense_bias; Type: Assign; Out #tensors: 1
Name: Placeholder; Type: Placeholder; Out #tensors: 1
Name: Conv2d; Type: Conv2D; Out #tensors: 1
Name: BiasAdd; Type: BiasAdd; Out #tensors: 1
Name: Activation_conv2d_1; Type: Relu; Out #tensors: 1
Name: Const_19; Type: Const; Out #tensors: 1
Name: Const_20; Type: Const; Out #tensors: 1
Name: MaxPool; Type: MaxPoolV2; Out #tensors: 1
Name: Conv2d_1; Type: Conv2D; Out #tensors: 1
Name: BiasAdd_1; Type: BiasAdd; Out #tensors: 1
Name: Activation_conv2d_3; Type: Relu; Out #tensors: 1
Name: Const_21; Type: Const; Out #tensors: 1
Name: Const_22; Type: Const; Out #tensors: 1
Name: MaxPool_1; Type: MaxPoolV2; Out #tensors: 1
Name: Reshape; Type: Reshape; Out #tensors: 1
Name: MatMul; Type: MatMul; Out #tensors: 1
Name: Add; Type: Add; Out #tensors: 1
Name: Activation_dense_6; Type: Relu; Out #tensors: 1
Name: MatMul_1; Type: MatMul; Out #tensors: 1
Name: Add_1; Type: Add; Out #tensors: 1
Name: SquaredDifference; Type: SquaredDifference; Out #tensors: 1
Name: Const_23; Type: Const; Out #tensors: 1
Name: Mean; Type: Mean; Out #tensors: 1
Name: Const_24; Type: Const; Out #tensors: 1
Name: default_training_loss; Type: Sum; Out #tensors: 1
Name: Gradients/OnesLike; Type: OnesLike; Out #tensors: 1
Name: Gradients/Shape; Type: Shape; Out #tensors: 1
Name: Gradients/Const; Type: Const; Out #tensors: 1
Name: Gradients/Const_1; Type: Const; Out #tensors: 1
Name: Gradients/Size; Type: Size; Out #tensors: 1
Name: Gradients/Add; Type: Add; Out #tensors: 1
Name: Gradients/Mod; Type: Mod; Out #tensors: 1
Name: Gradients/Range; Type: Range; Out #tensors: 1
Name: Gradients/OnesLike_1; Type: OnesLike; Out #tensors: 1
Name: Gradients/DynamicStitch; Type: DynamicStitch; Out #tensors: 1
Name: Gradients/Const_2; Type: Const; Out #tensors: 1
Name: Gradients/Maximum; Type: Maximum; Out #tensors: 1
Name: Gradients/Div; Type: Div; Out #tensors: 1
Name: Gradients/Reshape; Type: Reshape; Out #tensors: 1
Name: Gradients/Tile; Type: Tile; Out #tensors: 1
Name: Gradients/Shape_1; Type: Shape; Out #tensors: 1
Name: Gradients/Const_3; Type: Const; Out #tensors: 1
Name: Gradients/Const_4; Type: Const; Out #tensors: 1
Name: Gradients/Size_1; Type: Size; Out #tensors: 1
Name: Gradients/Add_1; Type: Add; Out #tensors: 1
Name: Gradients/Mod_1; Type: Mod; Out #tensors: 1
Name: Gradients/Range_1; Type: Range; Out #tensors: 1
Name: Gradients/OnesLike_2; Type: OnesLike; Out #tensors: 1
Name: Gradients/DynamicStitch_1; Type: DynamicStitch; Out #tensors: 1
Name: Gradients/Const_5; Type: Const; Out #tensors: 1
Name: Gradients/Maximum_1; Type: Maximum; Out #tensors: 1
Name: Gradients/Div_1; Type: Div; Out #tensors: 1
Name: Gradients/Reshape_1; Type: Reshape; Out #tensors: 1
Name: Gradients/Tile_1; Type: Tile; Out #tensors: 1
Name: Gradients/Shape_2; Type: Shape; Out #tensors: 1
Name: Gradients/Shape_3; Type: Shape; Out #tensors: 1
Name: Gradients/Const_6; Type: Const; Out #tensors: 1
Name: Gradients/Prod; Type: Prod; Out #tensors: 1
Name: Gradients/Prod_1; Type: Prod; Out #tensors: 1
Name: Gradients/Const_7; Type: Const; Out #tensors: 1
Name: Gradients/Maximum_2; Type: Maximum; Out #tensors: 1
Name: Gradients/Div_2; Type: Div; Out #tensors: 1
Name: Gradients/Cast; Type: Cast; Out #tensors: 1
Name: Gradients/Div_3; Type: Div; Out #tensors: 1
Name: Gradients/Const_8; Type: Const; Out #tensors: 1
Name: Gradients/Cast_1; Type: Cast; Out #tensors: 1
Name: Gradients/Subtract; Type: Sub; Out #tensors: 1
Name: Gradients/Multiply; Type: Mul; Out #tensors: 1
Name: Gradients/Multiply_1; Type: Mul; Out #tensors: 1
Name: Gradients/Negate; Type: Neg; Out #tensors: 1
Name: Gradients/Shape_4; Type: Shape; Out #tensors: 1
Name: Gradients/Shape_5; Type: Shape; Out #tensors: 1
Name: Gradients/BroadcastGradientArgs; Type: BroadcastGradientArgs; Out #tensors: 2
Name: Gradients/Sum; Type: Sum; Out #tensors: 1
Name: Gradients/Reshape_2; Type: Reshape; Out #tensors: 1
Name: Gradients/Sum_1; Type: Sum; Out #tensors: 1
Name: Gradients/Reshape_3; Type: Reshape; Out #tensors: 1
Name: Gradients/Identity; Type: Identity; Out #tensors: 1
Name: Gradients/Identity_1; Type: Identity; Out #tensors: 1
Name: Gradients/Shape_6; Type: Shape; Out #tensors: 1
Name: Gradients/Shape_7; Type: Shape; Out #tensors: 1
Name: Gradients/BroadcastGradientArgs_1; Type: BroadcastGradientArgs; Out #tensors: 2
Name: Gradients/Sum_2; Type: Sum; Out #tensors: 1
Name: Gradients/Reshape_4; Type: Reshape; Out #tensors: 1
Name: Gradients/Sum_3; Type: Sum; Out #tensors: 1
Name: Gradients/Reshape_5; Type: Reshape; Out #tensors: 1
Name: Gradients/MatMul; Type: MatMul; Out #tensors: 1
Name: Gradients/MatMul_1; Type: MatMul; Out #tensors: 1
Name: Gradients/ReluGrad; Type: ReluGrad; Out #tensors: 1
Name: Gradients/Identity_2; Type: Identity; Out #tensors: 1
Name: Gradients/Identity_3; Type: Identity; Out #tensors: 1
Name: Gradients/Shape_8; Type: Shape; Out #tensors: 1
Name: Gradients/Shape_9; Type: Shape; Out #tensors: 1
Name: Gradients/BroadcastGradientArgs_2; Type: BroadcastGradientArgs; Out #tensors: 2
Name: Gradients/Sum_4; Type: Sum; Out #tensors: 1
Name: Gradients/Reshape_6; Type: Reshape; Out #tensors: 1
Name: Gradients/Sum_5; Type: Sum; Out #tensors: 1
Name: Gradients/Reshape_7; Type: Reshape; Out #tensors: 1
Name: Gradients/MatMul_2; Type: MatMul; Out #tensors: 1
Name: Gradients/MatMul_3; Type: MatMul; Out #tensors: 1
Name: Gradients/Shape_10; Type: Shape; Out #tensors: 1
Name: Gradients/Reshape_8; Type: Reshape; Out #tensors: 1
Name: Gradients/MaxPoolGradV2; Type: MaxPoolGradV2; Out #tensors: 1
Name: Gradients/ReluGrad_1; Type: ReluGrad; Out #tensors: 1
Name: Gradients/BiasAddGrad; Type: BiasAddGrad; Out #tensors: 1
Name: Gradients/Identity_4; Type: Identity; Out #tensors: 1
Name: Gradients/Shape_11; Type: Shape; Out #tensors: 1
Name: Gradients/Conv2DBackpropInput; Type: Conv2DBackpropInput; Out #tensors: 1
Name: Gradients/Shape_12; Type: Shape; Out #tensors: 1
Name: Gradients/Conv2DBackpropFilter; Type: Conv2DBackpropFilter; Out #tensors: 1
Name: Gradients/MaxPoolGradV2_1; Type: MaxPoolGradV2; Out #tensors: 1
Name: Gradients/ReluGrad_2; Type: ReluGrad; Out #tensors: 1
Name: Gradients/BiasAddGrad_1; Type: BiasAddGrad; Out #tensors: 1
Name: Gradients/Identity_5; Type: Identity; Out #tensors: 1
Name: Gradients/Shape_13; Type: Shape; Out #tensors: 1
Name: Gradients/Conv2DBackpropInput_1; Type: Conv2DBackpropInput; Out #tensors: 1
Name: Gradients/Shape_14; Type: Shape; Out #tensors: 1
Name: Gradients/Conv2DBackpropFilter_1; Type: Conv2DBackpropFilter; Out #tensors: 1
Name: Const_25; Type: Const; Out #tensors: 1
Name: ApplyGradientDescent; Type: ApplyGradientDescent; Out #tensors: 1
Name: Const_26; Type: Const; Out #tensors: 1
Name: ApplyGradientDescent_1; Type: ApplyGradientDescent; Out #tensors: 1
Name: Const_27; Type: Const; Out #tensors: 1
Name: ApplyGradientDescent_2; Type: ApplyGradientDescent; Out #tensors: 1
Name: Const_28; Type: Const; Out #tensors: 1
Name: ApplyGradientDescent_3; Type: ApplyGradientDescent; Out #tensors: 1
Name: Const_29; Type: Const; Out #tensors: 1
Name: ApplyGradientDescent_4; Type: ApplyGradientDescent; Out #tensors: 1
Name: Const_30; Type: Const; Out #tensors: 1
Name: ApplyGradientDescent_5; Type: ApplyGradientDescent; Out #tensors: 1
Name: Const_31; Type: Const; Out #tensors: 1
Name: ApplyGradientDescent_6; Type: ApplyGradientDescent; Out #tensors: 1
Name: Const_32; Type: Const; Out #tensors: 1
Name: ApplyGradientDescent_7; Type: ApplyGradientDescent; Out #tensors: 1
21:49:39.814 [main] DEBUG api.keras.Sequential - Initialization of TensorFlow Graph variables
Train begins
Epoch 1 begins.
Training batch 0 begins.
21:49:40.380 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: 682.76715 metricValue: 0.102 }
Training batch 0 ends with loss 682.7671508789062.
Training batch 1 begins.
21:49:40.644 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: 463139.56 metricValue: 0.13 }
Training batch 1 ends with loss 463139.5625.
Training batch 2 begins.
21:49:40.909 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: 2.80750587E14 metricValue: 0.09 }
Training batch 2 ends with loss 2.80750586855424E14.
Training batch 3 begins.
21:49:41.154 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: Infinity metricValue: 0.094 }
Training batch 3 ends with loss Infinity.
Training batch 4 begins.
21:49:41.414 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.086 }
Training batch 4 ends with loss NaN.
Training batch 5 begins.
21:49:41.664 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.102 }
Training batch 5 ends with loss NaN.
Training batch 6 begins.
21:49:41.904 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 6 ends with loss NaN.
Training batch 7 begins.
21:49:42.167 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 7 ends with loss NaN.
Training batch 8 begins.
21:49:42.416 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 8 ends with loss NaN.
Training batch 9 begins.
21:49:42.687 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.102 }
Training batch 9 ends with loss NaN.
Training batch 10 begins.
21:49:42.954 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.124 }
Training batch 10 ends with loss NaN.
Training batch 11 begins.
21:49:43.194 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.102 }
Training batch 11 ends with loss NaN.
Training batch 12 begins.
21:49:43.424 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 12 ends with loss NaN.
Training batch 13 begins.
21:49:43.666 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.104 }
Training batch 13 ends with loss NaN.
Training batch 14 begins.
21:49:43.914 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.108 }
Training batch 14 ends with loss NaN.
Training batch 15 begins.
21:49:44.221 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.1 }
Training batch 15 ends with loss NaN.
Training batch 16 begins.
21:49:44.465 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.112 }
Training batch 16 ends with loss NaN.
Training batch 17 begins.
21:49:44.717 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.08 }
Training batch 17 ends with loss NaN.
Training batch 18 begins.
21:49:44.977 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.112 }
Training batch 18 ends with loss NaN.
Training batch 19 begins.
21:49:45.235 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.108 }
Training batch 19 ends with loss NaN.
Training batch 20 begins.
21:49:45.488 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.112 }
Training batch 20 ends with loss NaN.
Training batch 21 begins.
21:49:45.736 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.082 }
Training batch 21 ends with loss NaN.
Training batch 22 begins.
21:49:45.986 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.1 }
Training batch 22 ends with loss NaN.
Training batch 23 begins.
21:49:46.227 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.116 }
Training batch 23 ends with loss NaN.
Training batch 24 begins.
21:49:46.467 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.09 }
Training batch 24 ends with loss NaN.
Training batch 25 begins.
21:49:46.706 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 25 ends with loss NaN.
Training batch 26 begins.
21:49:46.943 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.114 }
Training batch 26 ends with loss NaN.
Training batch 27 begins.
21:49:47.189 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 27 ends with loss NaN.
Training batch 28 begins.
21:49:47.436 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 28 ends with loss NaN.
Training batch 29 begins.
21:49:47.687 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.086 }
Training batch 29 ends with loss NaN.
Training batch 30 begins.
21:49:47.927 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 30 ends with loss NaN.
Training batch 31 begins.
21:49:48.167 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.128 }
Training batch 31 ends with loss NaN.
Training batch 32 begins.
21:49:48.402 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 32 ends with loss NaN.
Training batch 33 begins.
21:49:48.642 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.08 }
Training batch 33 ends with loss NaN.
Training batch 34 begins.
21:49:48.882 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 34 ends with loss NaN.
Training batch 35 begins.
21:49:49.123 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.084 }
Training batch 35 ends with loss NaN.
Training batch 36 begins.
21:49:49.358 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.088 }
Training batch 36 ends with loss NaN.
Training batch 37 begins.
21:49:49.595 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.104 }
Training batch 37 ends with loss NaN.
Training batch 38 begins.
21:49:49.838 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.104 }
Training batch 38 ends with loss NaN.
Training batch 39 begins.
21:49:50.075 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.126 }
Training batch 39 ends with loss NaN.
Training batch 40 begins.
21:49:50.314 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.088 }
Training batch 40 ends with loss NaN.
Training batch 41 begins.
21:49:50.552 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.104 }
Training batch 41 ends with loss NaN.
Training batch 42 begins.
21:49:50.798 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.084 }
Training batch 42 ends with loss NaN.
Training batch 43 begins.
21:49:51.032 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.116 }
Training batch 43 ends with loss NaN.
Training batch 44 begins.
21:49:51.269 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 44 ends with loss NaN.
Training batch 45 begins.
21:49:51.502 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.102 }
Training batch 45 ends with loss NaN.
Training batch 46 begins.
21:49:51.739 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.088 }
Training batch 46 ends with loss NaN.
Training batch 47 begins.
21:49:51.977 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 47 ends with loss NaN.
Training batch 48 begins.
21:49:52.221 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.086 }
Training batch 48 ends with loss NaN.
Training batch 49 begins.
21:49:52.462 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 49 ends with loss NaN.
Training batch 50 begins.
21:49:52.702 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.088 }
Training batch 50 ends with loss NaN.
Training batch 51 begins.
21:49:52.948 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.09 }
Training batch 51 ends with loss NaN.
Training batch 52 begins.
21:49:53.189 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 52 ends with loss NaN.
Training batch 53 begins.
21:49:53.429 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.086 }
Training batch 53 ends with loss NaN.
Training batch 54 begins.
21:49:53.668 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.106 }
Training batch 54 ends with loss NaN.
Training batch 55 begins.
21:49:53.905 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.088 }
Training batch 55 ends with loss NaN.
Training batch 56 begins.
21:49:54.144 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.11 }
Training batch 56 ends with loss NaN.
Training batch 57 begins.
21:49:54.379 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.1 }
Training batch 57 ends with loss NaN.
Training batch 58 begins.
21:49:54.618 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.116 }
Training batch 58 ends with loss NaN.
Training batch 59 begins.
21:49:54.857 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 59 ends with loss NaN.
Training batch 60 begins.
21:49:55.097 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.114 }
Training batch 60 ends with loss NaN.
Training batch 61 begins.
21:49:55.351 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.086 }
Training batch 61 ends with loss NaN.
Training batch 62 begins.
21:49:55.610 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.106 }
Training batch 62 ends with loss NaN.
Training batch 63 begins.
21:49:55.855 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.07 }
Training batch 63 ends with loss NaN.
Training batch 64 begins.
21:49:56.101 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.082 }
Training batch 64 ends with loss NaN.
Training batch 65 begins.
21:49:56.346 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 65 ends with loss NaN.
Training batch 66 begins.
21:49:56.587 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.11 }
Training batch 66 ends with loss NaN.
Training batch 67 begins.
21:49:56.826 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 67 ends with loss NaN.
Training batch 68 begins.
21:49:57.062 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.102 }
Training batch 68 ends with loss NaN.
Training batch 69 begins.
21:49:57.306 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 69 ends with loss NaN.
Training batch 70 begins.
21:49:57.546 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.088 }
Training batch 70 ends with loss NaN.
Training batch 71 begins.
21:49:57.790 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.112 }
Training batch 71 ends with loss NaN.
Training batch 72 begins.
21:49:58.037 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 72 ends with loss NaN.
Training batch 73 begins.
21:49:58.285 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 73 ends with loss NaN.
Training batch 74 begins.
21:49:58.535 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.07 }
Training batch 74 ends with loss NaN.
Training batch 75 begins.
21:49:58.779 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.108 }
Training batch 75 ends with loss NaN.
Training batch 76 begins.
21:49:59.022 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.11 }
Training batch 76 ends with loss NaN.
Training batch 77 begins.
21:49:59.278 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 77 ends with loss NaN.
Training batch 78 begins.
21:49:59.518 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 78 ends with loss NaN.
Training batch 79 begins.
21:49:59.755 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.116 }
Training batch 79 ends with loss NaN.
Training batch 80 begins.
21:50:00.005 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.136 }
Training batch 80 ends with loss NaN.
Training batch 81 begins.
21:50:00.248 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.084 }
Training batch 81 ends with loss NaN.
Training batch 82 begins.
21:50:00.493 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.124 }
Training batch 82 ends with loss NaN.
Training batch 83 begins.
21:50:00.734 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 83 ends with loss NaN.
Training batch 84 begins.
21:50:00.982 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.1 }
Training batch 84 ends with loss NaN.
Training batch 85 begins.
21:50:01.222 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 85 ends with loss NaN.
Training batch 86 begins.
21:50:01.459 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 86 ends with loss NaN.
Training batch 87 begins.
21:50:01.700 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.074 }
Training batch 87 ends with loss NaN.
Training batch 88 begins.
21:50:01.955 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 88 ends with loss NaN.
Training batch 89 begins.
21:50:02.200 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.112 }
Training batch 89 ends with loss NaN.
Training batch 90 begins.
21:50:02.441 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 90 ends with loss NaN.
Training batch 91 begins.
21:50:02.676 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 91 ends with loss NaN.
Training batch 92 begins.
21:50:02.922 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.09 }
Training batch 92 ends with loss NaN.
Training batch 93 begins.
21:50:03.177 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.112 }
Training batch 93 ends with loss NaN.
Training batch 94 begins.
21:50:03.429 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.118 }
Training batch 94 ends with loss NaN.
Training batch 95 begins.
21:50:03.673 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.09 }
Training batch 95 ends with loss NaN.
Training batch 96 begins.
21:50:03.920 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 96 ends with loss NaN.
Training batch 97 begins.
21:50:04.165 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 97 ends with loss NaN.
Training batch 98 begins.
21:50:04.408 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.112 }
Training batch 98 ends with loss NaN.
Training batch 99 begins.
21:50:04.646 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.102 }
Training batch 99 ends with loss NaN.
Training batch 100 begins.
21:50:04.886 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.106 }
Training batch 100 ends with loss NaN.
Training batch 101 begins.
21:50:05.128 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 101 ends with loss NaN.
Training batch 102 begins.
21:50:05.376 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.11 }
Training batch 102 ends with loss NaN.
Training batch 103 begins.
21:50:05.625 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 103 ends with loss NaN.
Training batch 104 begins.
21:50:05.898 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.1 }
Training batch 104 ends with loss NaN.
Training batch 105 begins.
21:50:06.145 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 105 ends with loss NaN.
Training batch 106 begins.
21:50:06.384 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.11 }
Training batch 106 ends with loss NaN.
Training batch 107 begins.
21:50:06.646 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 107 ends with loss NaN.
Training batch 108 begins.
21:50:06.892 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 108 ends with loss NaN.
Training batch 109 begins.
21:50:07.145 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.104 }
Training batch 109 ends with loss NaN.
Training batch 110 begins.
21:50:07.395 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 110 ends with loss NaN.
Training batch 111 begins.
21:50:07.639 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 111 ends with loss NaN.
Training batch 112 begins.
21:50:07.877 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.094 }
Training batch 112 ends with loss NaN.
Training batch 113 begins.
21:50:08.121 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.098 }
Training batch 113 ends with loss NaN.
Training batch 114 begins.
21:50:08.363 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.104 }
Training batch 114 ends with loss NaN.
Training batch 115 begins.
21:50:08.605 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.102 }
Training batch 115 ends with loss NaN.
Training batch 116 begins.
21:50:08.845 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 116 ends with loss NaN.
Training batch 117 begins.
21:50:09.090 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.096 }
Training batch 117 ends with loss NaN.
Training batch 118 begins.
21:50:09.335 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.104 }
Training batch 118 ends with loss NaN.
Training batch 119 begins.
21:50:09.573 [main] DEBUG api.keras.Sequential - Batch stat: { lossValue: NaN metricValue: 0.092 }
Training batch 119 ends with loss NaN.
21:50:09.574 [main] INFO api.keras.Sequential - epochs: 1 loss: NaN metric: 0.098999985
Epoch 1 ends.
Train ends with last loss NaN
Test begins
Test batch 0 begins.
Test batch 0 ends with loss NaN..
Test batch 1 begins.
Test batch 1 ends with loss NaN..
Test batch 2 begins.
Test batch 2 ends with loss NaN..
Test batch 3 begins.
Test batch 3 ends with loss NaN..
Test batch 4 begins.
Test batch 4 ends with loss NaN..
Test batch 5 begins.
Test batch 5 ends with loss NaN..
Test batch 6 begins.
Test batch 6 ends with loss NaN..
Test batch 7 begins.
Test batch 7 ends with loss NaN..
Test batch 8 begins.
Test batch 8 ends with loss NaN..
Test batch 9 begins.
Test batch 9 ends with loss NaN..
Train ends with last loss NaN
Accuracy: 0.09800000488758087
from kotlindl.
@devcrocod Please, load and try a new version 0.0.9. This should be fixed.
from kotlindl.
I get new exception:
Exception in thread "main" java.lang.IllegalArgumentException: Negative dimension size caused by subtracting 2 from 1 for 'MaxPool_4' (op: 'MaxPoolV2') with input shapes: [?,1,1,128], [4], [4] and with computed input tensors: input[1] = <1 2 2 1>, input[2] = <1 2 2 1>.
at org.tensorflow.GraphOperationBuilder.finish(Native Method)
at org.tensorflow.GraphOperationBuilder.build(GraphOperationBuilder.java:42)
at org.tensorflow.GraphOperationBuilder.build(GraphOperationBuilder.java:21)
at org.tensorflow.op.nn.MaxPool.create(MaxPool.java:85)
at org.tensorflow.op.NnOps.maxPool(NnOps.java:621)
at api.core.layer.twodim.MaxPool2D.transformInput(MaxPool2D.kt:57)
at api.core.Sequential.transformInputWithNNModel(Sequential.kt:714)
at api.core.Sequential.compile(Sequential.kt:246)
at api.core.Sequential.compile(Sequential.kt:214)
at api.core.TrainableModel.compile$default(TrainableModel.kt:89)
at examples.keras.mnist.VGGMnistKt.main(VGGMnist.kt:175)
at examples.keras.mnist.VGGMnistKt.main(VGGMnist.kt)
from kotlindl.
from kotlindl.
Hmm, looks like you've got new exception on another example, VGGMnist.kt (at examples.keras.mnist.VGGMnistKt.main(VGGMnist.kt:175)), not on LeNetMnistWithCustomCallbacks
Please, check, that LeNetMnistWithCustomCallbacks has no problems during the run
from kotlindl.
@devcrocod Сould you please update your project with 0.0.10 version and check on your machine that this issue is fixed
from kotlindl.
LeNetMnistWithCustomCallbacks and VGGMnist run without exception. Loss function in LeNetMnistWithCustomCallbacks returns correct values.👍
from kotlindl.
from kotlindl.
Related Issues (20)
- Create tests for ONNX module for easy version upgrade HOT 1
- Visualization in example EfficientDetD2LightAPIwithVisualization.kt is broken HOT 3
- Upgrade Lets Plot For Swing/Batik to 3.1.0
- Upgrade Readme.md and Release Check List documentation HOT 1
- FloatArray to bitmap on Android HOT 1
- Add to documentation chapter in Readme links to the video/slides from 2023
- Dual head network like AlphaZero HOT 3
- Gradle Run Task Fails with SIGILL error in libtensorflow_framework.1.dylib HOT 1
- Documentation: create your first nn missing imports and dependencies, hard to follow HOT 3
- Upgrade ONNX version to 1.16.0
- Convert ONNX models to new opset version (at least 1.18)
- Upgrade kotlin-logging library HOT 1
- Add missed examples to test suites in examples module
- Support for Tensorflow 2.4 HOT 1
- Question
- Prepare a list of all supported models, hosted on S3
- TensorFlow problem: TensorFlow native library HOT 4
- Update Dataset type to support multilabel datasets HOT 1
- Why does drawObject accept coordinates of DetectedObject that are normalized? HOT 1
- Multiple outputs HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from kotlindl.