Tweeted By @jeremyphoward
Having spent a lot of time with both naive bayes and fine tuned neural nets, I thought this seemed wrong, so I ask the https://t.co/GEOZuodrZj community to check it out.
— Jeremy Howard (@jeremyphoward) June 6, 2019
At this stage, ULMFiT neural nets are winning the humor recognition competition: https://t.co/FJsEBbO6co https://t.co/YOQIXAZe1G