Gabriel Kissin

Firstly - much appreciation to @mwaskom for the awesome seaborn! A quick workaround which seems to work, exploiting the `units` parameter of `sns.lineplot`: ``` def lineplot_breaknans(data, break_at_nan=True, break_at_inf=True, **kwargs): '''sns.lineplot...

The problem has now been solved by managing to work with tensorforce 0.6.3

I've had this ZeroDivisionError in normal use, i.e. even when not restoring a population. The [docs](https://tfne.readthedocs.io/en/latest/codeepneat/codeepneat-specification.html?highlight=dynamic#module-speciation-type-param-distance-dynamic:~:text=If%20due%20to%20the%20random%20choice%20of%20modules%20for%20the%20blueprint%20graph%20an%20invalid%20TF%20model%20is%20generated%20from%20the%20genome%20genotype%2C%20the%20assembled%20genome%20is%20assigned%20a%20fitness%20score%20of%200) state "If due to the random choice of modules for the blueprint graph an...

Just to back up what I wrote in the previous comment, here is the output of the run which raised the error in Generation 46: ``` Evaluating 40 genomes in...

> I'm thinking of adding a speciation scheme that just uses k-means, so the only setting would be the number of species you want. The individuals would be grouped together...

Worth pointing out that the `tfne` library which implements the CoDeepNEAT algorithm (which is a successor to NEAT, for deep NNs) has functionality similar to what is being discussed here....

Thanks @vruusmann. Just spent a while working this out myself, came here to report it, and found this issue. The issue can also be seen if you replace `print(rulefit)` with...

related to https://github.com/csinva/imodels/issues/93, https://github.com/csinva/imodels/issues/77 - so this issue is a duplicate