Neural Turing Machine in Tensorflow

Tensorflow implementation of Neural Turing Machine. This implementation uses an LSTM controller. NTM models with multiple read/write heads are supported.

The referenced torch code can be found here.

* 1. Loss sometimes goes to nan even with the gradient clipping (#2).* * 2. The code is very poorly design to support NTM inputs with variable lengths. Just use this code as a reference.*

Prerequisites

• Python 2.7 or Python 3.3+

• Tensorflow 1.1.0

• NumPy

Usage

To train a copy task:

$ python main.py --task copy --is_train True

To test a quick copy task:

$ python main.py --task copy --test_max_length 10

Results

More detailed results can be found here.

Copy task:

Recall task:

(in progress)

Author

Taehoon Kim / @carpedm20