ibayer

Can you summarize again what you are trying to achieve? I saw you use mcmc somewhere in you code. Please keep in mind > It’s also possible to just call...

Looks to me like you use the mcmc solver. If that's the case then > I store w0,w_ , V_ from each run` doesn't make sense (I assume that by...

> None inform C++ side somehow so no cycles are wasted invoking the callback. I don't think invoking the callback will have any noticeable overhead if we do this ones...

I changed my mind a bit about passing parameter from cpp to python. Passing a string would be very flexible as this would allow us to serialize json. We already...

> The broad idea I had with such a cpp sends json string -> python parses json into dict is to have support ... I agree that's pretty cool. >...

Okay, I think I got a bit confused let's summarized it again. `python: call_back(string message) -> bool` message can also be used by cpp to send error message, bool decides...

Thanks for spelling out the details! Problem 1 > Should the solver stop because an error was found or continue pretending there is no issue in the callback while the...

> if we can add an attribute to feature mapping step as an intermediate step that has the same performance gains of this library? Can you add some details?

The paper you reference [0] addresses the cold start problem for positive only feedback recommender systems. The cold start problem is address with mapping functions many of them can be...

@jylim2016 Thanks for opening the PR. I think we need the following before merging: - The warm-start parameter should be added to https://github.com/ibayer/fastFM/blob/master/fastFM/sgd.py#L73 and https://github.com/ibayer/fastFM/blob/master/fastFM/sgd.py#L156 - A test case similar...