Hi, I would like to grab developers' attention to the issue that both "fit_vi" and "fit_hmc" functions are broken. For such a reason, unfortunately the CO_2 demo is not working.
While running "fit_hmc" in the notebook above, it triggers the following error message: TypeError: build_kernel.<locals>.kernel() got an unexpected keyword argument 'num_steps'
I have also tried replacing it with "fit_vi", but it stills triggers the following error message: AttributeError: 'PRNGKeyArray' object has no attribute 'init' in learning.py.
"fit_mle" is still working, though.
Lastly, the demo notebook about electricity is working fine. Nevertheless, I was wondering why setting an order>1 in sts.Autoregressive(order=1, ...) triggers an error. Perhaps it is a false alarm. Maybe, I am wrong since it is not mathematically possible, but I thought I could increase the order of the AR...
Hi, I would like to grab developers' attention to the issue that both
"fit_vi"and"fit_hmc"functions are broken. For such a reason, unfortunately the CO_2 demo is not working.While running
"fit_hmc"in the notebook above, it triggers the following error message:TypeError: build_kernel.<locals>.kernel() got an unexpected keyword argument 'num_steps'I have also tried replacing it with
"fit_vi", but it stills triggers the following error message:AttributeError: 'PRNGKeyArray' object has no attribute 'init'inlearning.py."fit_mle"is still working, though.Lastly, the demo notebook about electricity is working fine. Nevertheless, I was wondering why setting an
order>1insts.Autoregressive(order=1, ...)triggers an error. Perhaps it is a false alarm. Maybe, I am wrong since it is not mathematically possible, but I thought I could increase the order of the AR...