τ=0.5 errors

[SheldonCurtiss]

The error I was getting previously I've narrowed down to using
network_player_1 = Benchmark.NetworkOnly(τ=0.5)

when using
network_player_1 = Benchmark.NetworkOnly(τ=1.)

I don't have the error. What exactly does this control? Player temperature? Could this be caused due to my lack of normalizing my input data? Is input data normalized?

[jonathan-laurent]

Can you add the following line before the return statement of fix_probvec in util.jl:

@assert Distributions.isprobvec(π)

Then, if you run your example with τ=0.5 and it fails, can I see the stacktrace?

Also, you should make sure to always include all relevant context when filing an issue. You cannot expect me to spend more effort answering your questions than you spend writing them.

[SheldonCurtiss]

Apologies if it comes across that way, I was merely trying to get a description so I could further dig in myself.
I'm very appreciative of your responsiveness regarding my various questions.

I'll test and report back.

[jonathan-laurent]

I am closing this issue for lack of activity but please do not hesitate to reopen it if you need more help!