This page is just a springboard to the web domain I use to put my Deep/Machine Learning experiments in production: visualneurons.com.
It now offers users the possibility to:
- run Neural Style Transfer on images adopting the standard Gatys et al optimization approach (post).
- run the same as before on GIFs (post).
- run Neural Style Transfer on images adopting the novel, faster Johnson et al strategy, which involves training style models beforehand and no optimization at inference time (post).
- (feature disabled for the time being to save on the AWS bill) prompt a GPT-2 Huggingface language model and let it produce text (post).
- detect handwritten math and convert it to LaTeX syntax (post).
Below a screenshot from visualneurons.com, check it out!
