cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Highlighted
Contributor
Contributor
1,067 Views
Registered: ‎07-12-2019

[BACKEND][Check Failed: kernel_param * input_channel_group <= img_buf_depth]

Jump to solution

hello i am getting error from DNNC:

[DNNC][Warning] layer [activation_10_Softmax] (type: Softmax) is not supported in DPU, deploy it in CPU instead.
[BACKEND][Check Failed: kernel_param * input_channel_group <= img_buf_depth][/tmp/DNNDK_Pipeline_dnnc/dnnc/submodules/asicv2com/src/Operator/OperatorConv.cpp:75][DATA_OUTRANGE][Data value is out of range!] the kernel size is too much for current arch

this is my network in Keras :

model = Sequential()

model.add(Conv2D(filters=32, kernel_size=(3, 3), padding='same', input_shape=(224, 224, 3)))
model.add(Activation(activation=relu))
model.add(Conv2D(filters=32, kernel_size=(3, 3), padding='same'))
model.add(Activation(activation=relu))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(rate=0.2))

model.add(Conv2D(filters=64, kernel_size=(3, 3), padding='same'))
model.add(Activation(activation=relu))
model.add(Conv2D(filters=64, kernel_size=(3, 3), padding='same'))
model.add(Activation(activation=relu))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(rate=0.2))

model.add(Conv2D(filters=128, kernel_size=(3, 3), padding='same'))
model.add(Activation(activation=relu))
model.add(Conv2D(filters=128, kernel_size=(3, 3), padding='same'))
model.add(Activation(activation=relu))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(rate=0.2))

model.add(Conv2D(filters=256, kernel_size=(3, 3), padding='same'))
model.add(Activation(activation=relu))
model.add(Conv2D(filters=256, kernel_size=(3, 3), padding='same'))
model.add(Activation(activation=relu))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(rate=0.2))

model.add(Flatten())
model.add(Dense(units=256))
model.add(Activation(activation=relu))
model.add(Dropout(rate=0.7))
model.add(Dense(2))
model.add(Activation(activation=softmax))


0 Kudos
1 Solution

Accepted Solutions
Highlighted
Contributor
Contributor
1,053 Views
Registered: ‎07-12-2019
0 Kudos
4 Replies
Highlighted
Contributor
Contributor
1,054 Views
Registered: ‎07-12-2019
0 Kudos
Highlighted
Observer
Observer
934 Views
Registered: ‎07-19-2018

Hi, I got the same error with Keras model mobilenet_v2. Could you tell me how did you fix it?

Thanks a lot!

Regards,

dvukadin

0 Kudos
Highlighted
Contributor
Contributor
923 Views
Registered: ‎07-12-2019

This error occurs when you have to many neurons enetering Dense layer. It depends which board you are using. The formula is Input_channel <= 2048 * channel_parallel. You can check it in DPU for Convolutional Neural Network v3.0 at page 21. I am using ZCU102 so i my channel_parrarel is equal 16 which means i can have 32768 neurons enetering Dense layer but i had 14x14x256 which was 50176. You have to reduce number of neuron entering Dense layer

 

Highlighted
Observer
Observer
862 Views
Registered: ‎07-19-2018
Thanks!
0 Kudos