search

aymericdamien / TensorFlow-Examples

TensorFlow Tutorial and Examples for Beginners (support TF v1 & v2)
Other
43.38k stars 14.95k forks source link

Understanding Usage of Timesteps #251

Open raagas opened 6 years ago

raagas commented 6 years ago

Hi, Thanks for putting this wonderful tutorial. For beginner of tensorflow, this is extremely helpful.

In RNN, one thing I'm not understanding one thing. Why we are using timesteps in the following placeholder? X = tf.placeholder("float", [None, timesteps, num_input])

We could just have used: X = tf.placeholder("float", [None, num_steps]),

where None represents the bact_size or data_size and num_steps represents the "how far RNN is looking into past". Similarly, I'm not getting the usage of num_input.

It will be great if you help me understand the parameters. Or can guide me to some tutorial, which explains usage of timesteps and num_inputs.

Thanks!

mmuratarat commented 6 years ago

Because RNNs require sequences.