May 2018
Beginner
490 pages
13h 16m
English
A GAN can be used to generate the large datasets required for CRLMMs. The real data could be a set of images with the correct word vectors. The generator could then produce millions of images and labels that the discriminator would work on. The accurately generated datasets could be stored in a CRL repository.
A GAN can be implemented in a few lines of code in TensorFlow, for example. The Further reading section at the end of the chapter provides links that detail this simple but effective type of network.
Read now
Unlock full access