MXNet (using Gluon) implementation of End-To-End Memory Networks for language modelling. The original Tensorflow code from carpedm20 can be found here. MXNet Symbolic implementation can be found here.
$ pip install -r requirements.txt
To train a model with 6 hops and memory size of 100, run the following command:
$ python main.py --nhop 6 --mem_size 100
To test a model with the lastest stored model:
$ python main.py --is_test True
To see all options, run:
$ python main.py --help