The meaning of "number of training example" ? "

Hi guys,

I am in the “Improvise_a_Jazz_Solo_with_an_LSTM_Network_v4”

I want to ask, what’s the meaning of “number of training examples: 60” for the mid file ?

So, I opened this mid file, it’s ONE song. It looks like One example only. This 60 means there are 60 different “instruments” playing together for this ONE song? And LSTM is taking one song which is played by one instrument one at a time?
Thank you!

Hi @sunson29 ,

Please refer to the mark-up cell just below the output of these print statements. Those numbers are explained in details.

Yes, I read that.

  1. For the m, it just says "You have m training examples, ". I am asking, what’s the physically meaning of 60 in ONE song?

  2. For the Tx=30, my understanding is there are 30 total notes played for each example.

Hi @sunson29 ,

The physical meaning of 60 training examples from this piece of Jazz music is referring to 60 sequence pieces of musical values, each sequence piece is consists of 30 values. The comment lines explained that these musical values can be consider as notes, but in music theory, the term value is more complicated, and specifically, it also captures the information needed to play multiple notes at the same time.

For this assignment, the data pre-processing is done for us by a couple of utility functions which extract the signature elements of Jazz music from this one piece of music and returned the training dataset X. This dataset X consists of sequence pieces from various instruments picked up from this piece of Jazz music. The caveat here is that I don’t know much about how music works and I don’t have contact with the author of this assignment.

If you would like to take a look at how these utility functions work:
file ->open->
data_util.pys
preprocess.py
music_util.py

1 Like

thank you very much!