Text2Image generation
How did I implement it ?
Using code similar to the one provided by VQGAN+CLIP or some Colab. Added also a small Perceptual loss to increase the resemblance with an input image. The code has been pushed to github.
Other possibilities
Diffusion with CLIP Deep daze Big Sleep With image input
Some examples
A calm chalet in the Alps: painting inspiration.
Computer vision.
A mystery man walking in the streets.
A man in the forest by Picasso.
A view of the sea by Gauguin
Ugo Tanielian
My research interests include deep learning theory. In particular, I am very interested in both generative models and time series. My thesis dealt with Generative Adversarial Networks.