Problem of creating lstm input

[jeremy chang]

Hi, I just started to use jeffdonahue  's lstm recently

I want to ask the setting of input data and prototxt

i want to use lstm to predict a sin wave

f(x)=sin(x)

my time step input(x) is like

1   2   3   4   5   6   7   8   9   10
11 12 13 14 15 16 17 18 19 20

indicator is like

0  1    1   1   1   1   1   1   1    1    

0  1    1   1   1   1   1   1   1    1    

meaning i want to take the 2 rows as independent input sequence

output is like

sin(1)   sin(2)   sin(3)   sin(4)   sin(5)   sin(6)   sin(7)   sin(8)   sin(9)    sin(10)

sin(11) sin(12) sin(13) sin(14) sin(15) sin(16) sin(17) sin(18) sin(19)  sin(20)

all of the data above in saved in a hdf5 file

and my prototxt is like

 

name: "sin_predict_model"

layer {

    name: "data"

type: "HDF5Data"

top: "input"

top: "output"

top: "indicator"

include {phase:TRAIN}

hdf5_data_param {

  source:"./sin.txt"

  batch_size:5

}

}

i thought this setting should be like

sending 5 (batch_size) independent sequences , each have 10(10 elements in a row)  time step

but this message appeared

"Initializing recurrent layer: assuming input batch contains 5 timesteps of 10 independent streams.

"

Did i misunderstand anything?

what should the correct setting be like?

thanks!