How to load and lstm forecast several excel columns at a time

I have an excel file with 3,000 rows where each row represents a product with its monthly sales in the columns (about 5 years worth of monthly sales).

I have gone through a tutorial on how to make an LSTM forecast.
I have seen some videos on how to import selected rows.
I learned how to use np.savetxt(‘107084.csv’, df_proj, delimiter=’,’)

But what I need is to select every row in the excel file individually, create a forecast for each and put the forecast into a new excel file or into individual csv files that i can later merge into a single file.

Also the exported data is this:
nan,-2.399401724338531494e+00
nan,7.489869117736816406e+00
nan,1.482942080497741699e+01
nan,2.937255978584289551e+01
nan,2.080834364891052246e+01
nan,2.041447734832763672e+01
nan,1.649583864212036133e+01
nan,2.451917123794555664e+01
nan,2.374986648559570312e+01
nan,1.717235398292541504e+01
nan,5.879066884517669678e+00
nan,2.848649954795837402e+01

without the month column.

I need guidance please.