Oren Matar

Hi, @tcuongd, @winedarksea, are there any updates or expected timeline for merging and releasing this update? Is there anything I can do to help speed it up?

@nicolaerosia @winedarksea that's strange, I'd expect a much greater speed improvement based on: https://towardsdatascience.com/how-to-run-facebook-prophet-predict-x100-faster-cce0282ca77d

@tcuongd I went over the code and it looks really well integrated. My only comment is the names of the vectorized functions are not very indicative - the same name...

@winedarksea I don't know of any attempt... and would love to see one. If you have any questions about the code - I'd be happy to help. BTW, I'm working...

@winedarksea 1. that's right - k is just the number of samples we take for the trend 2. As far as I could tell from the code it doesn't influence...

@winedarksea How is it going with the project? BTW, I just published another post on optimizing Prophet, which you may find interesting... https://medium.com/towards-data-science/need-for-speed-optimizing-facebook-prophet-fit-method-to-run-20x-faster-166cd258f456

@winedarksea I think 60ms is slow compared to most linear regressions which is what Prophet ultimately is. some datasets really do contain hundreds of thousands of items if not millions,...

@nicolaerosia I don't... I only implemented it as a side function to be called separately and @winedarksea started working on a fork. Not too difficult to replace the existing confidence...

@tcuongd @winedarksea @nicolaerosia Finally had some time to check why this version was still significantly slower than what I got in my original blog post. The culprit: DataFrames. Their init...