Decomposing a complex, state-of-the-art model implementation
Even though you have picked up the basics of TF and PyTorch, setting up a model training from scratch can be overwhelming. Luckily, the two frameworks have thorough documentations and tutorials that are easy to follow:
- TF
- Image classification with convolution layers: https://www.tensorflow.org/tutorials/images/classification.
- Text classification with recurrent layers: https://www.tensorflow.org/text/tutorials/text_classification_rnn.
- PyTorch
- Object detection with convolutional layers: https://pytorch.org/tutorials/intermediate/torchvision_tutorial.html.
- Machine translation with recurrent layers: https://pytorch.org/tutorials/intermediate/seq2seq_translation_tutorial.html.
In this section, we would like to look at a model that is much more sophisticated, StyleGAN. Our main goal is to explain how the components described in the previous sections can be put together for a complex DL project. For the complete description...
