Using Deep Learning for sequencial data

I have a big number of sequences of labeled Data and I would like to use Deep Learning to classify them. For example: I have n files (f1 f2 f3…fn) Each one has a very large number of sequences:

f1

0 15 0 0 0 2…

0 2 4 4 4 8…

1 0 0 5 7 8…

0 3 5 6 0 22…

f2

0 0 0 0 1 0…

1 0 3 0 5 0…

3 0 0 0 11 4…

2 0 7 8 1 3…

Now for a given sequence (e.g: 1 4 3 0 0 0 0…), the (deep learning) program should be able to classify it in the correct file. Could you explain in easy_to_understand way how to tackle this problem using deep learning ? I am planning to use R and/or Python. I will be very glad to hear about some useful deep learning libraries for such problems.

Can i Do Like this,

Try to use LSTM for this task. Sequential task is a natural use of LSTM. You can use python, has several libraries such as theano, pybrain, tensorflow, torch, etc. For example, sound is a sequential phenomenon and modern speech recognition(google,microsoft,baidu) use LST for this task.