def test_torch_rnn_classifier_cheese_disease(cheese_disease_dataset):
mod = torch_rnn_classifier.TorchRNNClassifier(
vocab=cheese_disease_dataset['vocab'],
embed_dim=20,
hidden_dim=20,
max_iter=20)
mod.fit(cheese_disease_dataset['X_train'], cheese_disease_dataset['y_train'])
pred = mod.predict(cheese_disease_dataset['X_test'])
assert accuracy_score(cheese_disease_dataset['y_test'], pred) > 0.80
def test_torch_rnn_classifier_incremental(X_sequence):
train, test, vocab = X_sequence
model = torch_rnn_classifier.TorchRNNClassifier(vocab=vocab, max_iter=100)
X, y = zip(*train)
X_test, _ = zip(*test)
model.fit(X, y, X_dev=X_test, dev_iter=20)
epochs = list(model.dev_predictions.keys())
assert epochs == list(range(20, 101, 20))
assert all(len(v) == len(X_test) for v in model.dev_predictions.values())
def test_torch_rnn_classifier(X_sequence):
"""Just makes sure that this code will run; it doesn't check that
it is creating good models.
"""
train, test, vocab = X_sequence
mod = torch_rnn_classifier.TorchRNNClassifier(vocab=vocab, max_iter=100)
X, y = zip(*train)
X_test, _ = zip(*test)
mod.fit(X, y)
mod.predict(X_test)
mod.predict_proba(X_test)
def test_torch_rnn_classifier_save_load(X_sequence):
train, test, vocab = X_sequence
mod = torch_rnn_classifier.TorchRNNClassifier(vocab=vocab, max_iter=2)
X, y = zip(*train)
X_test, _ = zip(*test)
mod.fit(X, y)
mod.predict(X)
with tempfile.NamedTemporaryFile(mode='wb') as f:
name = f.name
mod.to_pickle(name)
mod2 = torch_rnn_classifier.TorchRNNClassifier.from_pickle(name)
mod2.predict(X_test)
mod2.fit(X, y)
acc = np_sgd_classifier.simple_example()
assert acc >= 0.89
@pytest.mark.parametrize("model, params", [
[
np_sgd_classifier.BasicSGDClassifier(max_iter=10, eta=0.1),
{'max_iter': 100, 'eta': 1.0}
],
[
np_rnn_classifier.RNNClassifier(
vocab=[], max_iter=10, hidden_dim=5, eta=0.1),
{'hidden_dim': 10, 'eta': 1.0, 'max_iter': 100}
],
[
torch_rnn_classifier.TorchRNNClassifier(
vocab=[], max_iter=10, hidden_dim=5, eta=0.1),
{
'hidden_dim': 10,
'eta': 1.0,
'max_iter': 100,
'l2_strength': 0.01,
'embed_dim': 100,
'bidirectional': False
}
],
[
np_tree_nn.TreeNN(
vocab=[], max_iter=10, hidden_dim=5, eta=0.1),
{'embed_dim': 5, 'hidden_dim': 10, 'eta': 1.0, 'max_iter': 100}
],
[
