ValueError: sum(pvals[:-1]) > 1.0

[saitaiky]

def __sample(a, temperature=1.0):
a = np.log(a) / temperature
a = np.exp(a) / np.sum(np.exp(a))
return np.argmax(np.random.multinomial(1, a, 1))

It shows up multinomial has an error
np.argmax(np.random.multinomial(1, a, 1))
File "mtrand.pyx", line 4593, in mtrand.RandomState.multinomial (numpy\random\mtrand\mtrand.c:37541)
ValueError: sum(pvals[:-1]) > 1.0

[saitaiky]

Here is my solution: The problem is because of the probabilities sometimes sum to just above 1.0
I dunno much about the maths, but there is my work around.
a = np.log(a) / temperature dist = np.exp(a)/np.sum(np.exp(a)) choices = range(len(a)) return np.random.choice(choices, p=dist)

[developerator]

Thanks a lot for this workaround, had the same issue. @llSourcell Please adopt it

[japrogramer]

Had this problem too, @saitai0802 patch seems to allow it to continue and finish. not shure if the result is better, if the probabilities are adding up to more than 1, something else is causing that.

[jasondesante]

I have the same issue

was fixed with the workaround

then I got another error that said I needed to install pygame

after that it worked

but I also had a bunch of errors before posting here, like a whole bunch of the code was written for python 2 but I'm using python 3.....not sure if I was supposed to use python 2 at the start

[PandaWhoCodes]

I have uploaded a python3 version here - for anyone who needs it