Thanks @cluhmann! Yes these are copy and past errors. I was trying to do do tt.stack on about 18000 data frames in dic_data each with a date time index of 12 months (each data frame needs to be processed separately for time lag within the data frame). I think my windows machine could not handle it. So I reduced the number of data frames in dic_data to 2, 10, 100 respectively and the code now runs. My thinking is that, on my windows laptop, the theano computational engine is unable to create a computational graph for such a large amount of data frames? I am new to theano. Thanks
1 Like