No module named 'pykg2vec.core.SMEBilinear'

[tobimichigan]

Issue Template

First of all, thank you for finding the bugs. Please help us improve by providing the bug report as much as possible in the following format.

Describe the bug Tried running one of your notebooks

To Reproduce `# Author: Sujit Rokka Chhetri and Shiy Yuan Yu

License: MIT

import tensorflow as tf

from pykg2vec.core.TransE import TransE from pykg2vec.core.TransH import TransH from pykg2vec.core.TransR import TransR from pykg2vec.core.Rescal import Rescal from pykg2vec.core.SMEBilinear import SMEBilinear from pykg2vec.core.SMELinear import SMELinear from pykg2vec.config.config import TransEConfig, TransHConfig, TransRConfig, RescalConfig, SMEConfig

from pykg2vec.utils.dataprep import DataPrep from pykg2vec.utils.trainer import Trainer

def experiment():

preparing dataset.

knowledge_graph = DataPrep('Freebase15k')

# preparing settings. 
epochs = 5
batch_size = 128
learning_rate = 0.01
hidden_size = 50

transEconfig = TransEConfig(learning_rate=learning_rate,
                            batch_size=batch_size,
                            epochs=epochs, hidden_size=hidden_size)

transHconfig = TransHConfig(learning_rate=learning_rate,
                            batch_size=batch_size,
                            epochs=epochs, hidden_size=hidden_size)

transRconfig = TransRConfig(learning_rate=learning_rate,
                            batch_size=batch_size,
                            ent_hidden_size=64,
                            rel_hidden_size=32,
                            epochs=epochs)

rescalconfig = RescalConfig(learning_rate=0.1,
                            batch_size=batch_size,
                            epochs=epochs, hidden_size=hidden_size)

smeconfig = SMEConfig(learning_rate=learning_rate,
                      batch_size=batch_size,
                      epochs=epochs, hidden_size=hidden_size)

configs = [transEconfig, transHconfig, transRconfig, rescalconfig, smeconfig]

for config in configs:
    config.test_step = 2
    config.test_num = 100
    config.save_model = True
    config.disp_result = False

# preparing models. 
models = []
models.append(TransE(transEconfig, knowledge_graph))
models.append(TransH(transHconfig, knowledge_graph))
models.append(TransR(transRconfig, knowledge_graph))
models.append(Rescal(rescalconfig, knowledge_graph))
models.append(SMEBilinear(smeconfig, knowledge_graph))
models.append(SMELinear(smeconfig, knowledge_graph))

# train models.
for model in models:
    print("training model %s" % model.model_name)
    trainer = Trainer(model=model)

    trainer.build_model()
    trainer.train_model()
    trainer.full_test()

    tf.reset_default_graph()

if name == "main": experiment()`

Expected behavior Perfect run

Screenshots If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information): Linux

• SDK Conda

Dependency (please complete the following information):

• Current version of the dependencies tensorflow>=2.1.0 sphinx>=2.1.2 networkx>=2.2 setuptools>=40.8.0 matplotlib>=3.0.3 numpy>=1.16.2 seaborn>=0.9.0 scikit_learn>=0.20.3 hyperopt>=0.2.1 pathlib>=1.0.1 numpydoc>=0.9.1 sphinx-gallery>=0.3.1 sphinx-rtd-theme>=0.4.3 pytest>=3.6 Additional context Any clues?

[louisccc]

Hello, thank you for reporting this issue.

The notebook is kind of out-of-date. We will work on revising them asap.

For now, if you want to try training KGE, I will suggest you run the scripts in [root folder]/pykg2vec/example/train.py. and by modifying argument followed by "-mn", you can toggle between models.!

[tobimichigan]

Thanks again @louisccc . It seems that needs to be done quick;y cos its affecting the entire notebook SDK.

[louisccc]

We have updated the example scripts so you may want to check them out under [root]/examples. We are still actively updating the user documents. so you can expect a better version to be coming up in a short while! Thanks!