ibayer

> Is the maximum possible n_iter related to the size of the training data? No, n_iter is really the number of individual sgd updates (not epoch). This is indeed the...

Thank you for your kind words. A contributions removing the assertions would indeed be very welcome. A good place to start would be this issue: https://github.com/ibayer/fastFM/issues/88

@chunguo Thanks

@PyPatel Your are on CentOS 7 as well? fastFM is only tested on ubuntu but cross compiled binaries should work across different distributions.

Okay, I think we have tested on ubuntu 14.04 before. Can you create a fully reproducible example incl. Python & library versions?

Hyper parameter tuning is more art then science, make sure that you also tune, rank, n_iter and even init_std (not all are equally important). 1. The fact that MCMC has...

1. Use what ever gives you the best result. 2. link updated 3. Go with ALS, MCMC doesn't really make sense for user / item embeddings (you would get a...

> 2) This seems a bit in conflict with the answer to Question 4 where ALS is suggested as superior in this situation. No, I didn't say **superior** " (you...

@ogencoglu Thanks for the feedback 3.8 is currently not supported. We are working on a complete rewrite which will bring 3.8 support and much more.

@mohelm We made good progress but currently all free resources go into a public service covid19 related project. Feel free to ping me again in a few week, I should...