<< Click to Display Table of Contents >> Navigation:  Introduction > Architectures included in Distribution > ResNet50

ResNet-50 is written in TensorFlow, the architecture contains residual blocks and 50 layers.

The network is dedicated for classification. In PAI, the available ResNet-50 architecture is adapted for 2D and 3D data. This architecture is available as ResNet 50 in the Create Learning Set dialog. The loss function used for this architecture is categorical cross entropy.

 

References

[1] Keras repo, ResNet50 implementation

https://github.com/keras-team/keras-applications/blob/master/keras_applications/resnet50.py

[2] Priya Dwivedi, towards data science

https://towardsdatascience.com/understanding-and-coding-a-resnet-in-keras-446d7ff84d33

https://github.com/priya-dwivedi/Deep-Learning/blob/master/resnet_keras/Residual_Networks_yourself.ipynb