TensorFlow implementation of Pointer Networks.
(in progress)
- Python 2.7
- tqdm
- TensorFlow 0.12.1
To train a model:
$ python main.py --task=tsp --max_data_length=20 # download dataset from the paper $ python main.py --task=tsp --max_data_length=10 # generate dataset itself To train a model:
$ python main.py $ tensorboard --logdir=logs --host=0.0.0.0 To test a model:
$ python main.py --is_train=False Train/Test loss of max_data_length=10 after 40,000 steps:
$ python main.py --reg_scale=1.0 --optimizer=sgd (in progress)
Taehoon Kim / @carpedm20