Is the argument X of the predict function a matrix of 60 time-series, with length 200, from all the 6 categories? Is it, for example, 10 time-series for each category? Moreover, do we need to consider the 6 categories as a single input feature/variable? Thanks
Posted by: elisa_brosera @ Dec. 7, 2023, 4:17 p.m.There are 10 time series for each category. It is not mandatory to use the category variable, is just an additional information.
Posted by: an2dl.competitions @ Dec. 7, 2023, 5:05 p.m.Is, then, the shape of X (BS=60, 200,1)? Thanks
Posted by: elisa_brosera @ Dec. 7, 2023, 5:23 p.m.I did my model with an input shape of (60,200) but when i'm submitting it, i receive the error "ValueError: Input 0 of layer "CONV_LSTM_model" is incompatible with the layer: expected shape=(None, 60, 200), found shape=(None, 200)".
Posted by: jack_sansoni @ Dec. 10, 2023, 9:42 p.m.The input array has two dimensions with shape (60, 200).
Posted by: an2dl.competitions @ Dec. 11, 2023, 1:31 p.m.