3 ms·
It is not a full course but, since the fast.ai library does cover tabular data[1], they touch the subject in lesson 4[2]. [1]: https://docs.fast.ai/tabular.htm
by nestorD 7y ago
It is not a full course but, since the fast.ai library does cover tabular data[1], they touch the subject in lesson 4[2].
[1]: https://docs.fast.ai/tabular.html https://docs.fast.ai/tabular.html
[2]: https://course.fast.ai/videos/?lesson=4 https://course.fast.ai/videos/?lesson=4
- jph00 7y agoYou might also be interested in the sister course to the one posted here: "Introduction to Machine Learning for Coders": http://course18.fast.ai/ml http://course18.fast.ai/ml
- kashprime 7y agoJust a question for a project of mine, but can tabular data be combined with NLP? Or should I train two seperate nets (i.e. with ulmfit) and try to combine the results?
- victor9000 7y agoYou can one-hot encode categorical data like strings to build one big model. But if the data is independent, or if the tasks are orthogonal then you might be better served if you create two models.