I seen tf.nn.dynamic_rnn in tensorflow documentation how do I prepare some example data csv and python time sereis info and full code?
josheeg opened this issue · comments
I seen tf.nn.dynamic_rnn in tensorflow documentation how do I prepare some example data csv and python time sereis info and full code?
I feel there is not enough information to use this for me.
More generally, the DNC class found within dnc.py can be used as a standard TensorFlow rnn core and unrolled with TensorFlow rnn ops, such as tf.nn.dynamic_rnn on any sequential task.
You can use tf.TextLineReader to read in CSV files to TensorFlow, more details here:
https://www.tensorflow.org/api_guides/python/reading_data
E.g. run_model in train.py just needs to have tensor inputs of shape [sequence_length, batch_size, input_size]. This is just an example script, more generally you can use the DNC class on any sequential task and use TensorFlow loop primitives (such as dynamic_rnn or static_rnn) to unroll the DNC and perform BPTT.