def test_predict_functions_honor_device(digits, func):
X_train, X_test, y_train, y_test = digits
mod = TorchShallowNeuralClassifier(max_iter=2)
mod.fit(X_train, y_train)
prediction_func = getattr(mod, func)
with pytest.raises(RuntimeError):
prediction_func(X_test, device="FAKE_DEVICE")
def test_hidden_activation_in_graph(digits):
X_train, X_test, y_train, y_test = digits
mod = TorchShallowNeuralClassifier(max_iter=1, hidden_activation=nn.ReLU())
mod.fit(X_train, y_train)
mod_hidden_activation = mod.hidden_activation.__class__
graph_activation_class = mod.model[1].__class__
assert mod_hidden_activation == graph_activation_class
def test_model_graph_dimensions(digits, attr, layer_index, weight_dim):
X_train, X_test, y_train, y_test = digits
mod = TorchShallowNeuralClassifier(max_iter=1)
mod.fit(X_train, y_train)
mod_attr_val = getattr(mod, attr)
graph_dim = mod.model[layer_index].weight.shape[weight_dim]
assert mod_attr_val == graph_dim
def test_simple_example_params(digits, param, expected):
X_train, X_test, y_train, y_test = digits
mod = TorchShallowNeuralClassifier(**{param: expected})
mod.fit(X_train, y_train)
preds = mod.predict(X_test)
acc = accuracy_score(y_test, preds)
if not (param in ["max_iter", "batch_size"] and expected <= 1):
assert acc >= 0.86
def test_predict_restores_device(digits, func):
X_train, X_test, y_train, y_test = digits
mod = TorchShallowNeuralClassifier(max_iter=2)
mod.fit(X_train, y_train)
current_device = mod.device
assert current_device != torch.device("cpu:0")
prediction_func = getattr(mod, func)
prediction_func(X_test, device="cpu:0")
assert mod.device == current_device
def test_model(XOR):
"""Just makes sure that this code will run; it doesn't check that
it is creating good models.
"""
X, y = XOR
model = TorchShallowNeuralClassifier(hidden_dim=4,
hidden_activation=nn.ReLU(),
max_iter=100,
eta=0.01)
model.fit(X, y)
model.predict(X)
model.predict_proba(X)
def fit_basic_rnn(X, y, hidden_dim, max_iter, hidden_activation, eta): if hidden_dim is None: hidden_dim = 50 if max_iter is None: max_iter = 100 if hidden_activation is None: hidden_activation = nn.Tanh() if eta is None: eta = 0.01 mod = TorchShallowNeuralClassifier(hidden_dim = hidden_dim, max_iter = max_iter, hidden_activation = hidden_activation, eta = eta) mod.fit(X, y) return mod
def test_save_load(XOR):
X, y = XOR
mod = TorchShallowNeuralClassifier(hidden_dim=4,
hidden_activation=nn.ReLU(),
max_iter=100,
eta=0.01)
mod.fit(X, y)
mod.predict(X)
with tempfile.NamedTemporaryFile(mode='wb') as f:
name = f.name
mod.to_pickle(name)
mod2 = TorchShallowNeuralClassifier.from_pickle(name)
mod2.predict(X)
mod2.fit(X, y)
def test_optimizer_keywords(XOR, expected):
X, y = XOR
mod = TorchShallowNeuralClassifier(amsgrad=expected)
mod.fit(X, y)
assert mod.amsgrad == expected
assert mod.optimizer.param_groups[0]['amsgrad'] == expected
def fit_hf_shallow_network(X, y):
mod = TorchShallowNeuralClassifier(
max_iter=100, hidden_dim=300)
mod.fit(X, y)
return mod
def fit_nn_classifier(X, y):
mod = TorchShallowNeuralClassifier(hidden_dim=50, max_iter=100)
mod.fit(X, y)
return mod
class Shallow_Neural_Classifier:
'''
Modified torch shallow neural classifier wrapper class for initial fitting and then fine tuning of weights.
'''
def __init__(self, sent140_train_X, sent140_dev_X, sent140_train_Y, sent140_dev_Y, emoji_train_X, emoji_dev_X, emoji_test_X, emoji_train_Y, emoji_dev_Y, emoji_test_Y, emojiless_train_X, emojiless_dev_X, emojiless_test_X, emojiless_train_Y, emojiless_dev_Y, emojiless_test_Y, testing):
'''
Pass in initial data for fitting to constructor. Later adding passing logisitic regression
parameters into constructor.
'''
self.testing = testing
self.sent140_train_X = sent140_train_X
self.sent140_train_Y = sent140_train_Y
self.sent140_dev_X = sent140_dev_X
self.sent140_dev_Y = sent140_dev_Y
self.emoji_train_X = emoji_train_X
self.emoji_train_Y = emoji_train_Y
self.emoji_dev_X = emoji_dev_X
self.emoji_dev_Y = emoji_dev_Y
if self.testing:
self.emoji_test_X = emoji_test_X
self.emoji_test_Y = emoji_test_Y
self.emojiless_train_X = emojiless_train_X
self.emojiless_train_Y = emojiless_train_Y
self.emojiless_dev_X = emojiless_dev_X
self.emojiless_dev_Y = emojiless_dev_Y
if self.testing:
self.emojiless_test_X = emojiless_test_X
self.emojiless_test_Y = emojiless_test_Y
# pass in model parameters for to constructor?
def run_sent140(self):
'''
Trained on sent140, predict on emoji
Report score on sent 140 too, just because it's intersting
'''
# model
self.model_sent140 = TorchShallowNeuralClassifier()
# train
self.model_sent140.fit(self.sent140_train_X, self.sent140_train_Y)
# test on sent140
sent140_train_preds = self.model_sent140.predict(self.sent140_train_X)
sent140_dev_preds = self.model_sent140.predict(self.sent140_dev_X)
# test on emoji
emoji_train_preds = self.model_sent140.predict(self.emoji_train_X)
emoji_dev_preds = self.model_sent140.predict(self.emoji_dev_X)
if self.testing:
emoji_test_preds = self.model_sent140.predict(self.emoji_test_X)
else:
emoji_test_preds = None
return (sent140_train_preds, sent140_dev_preds, emoji_train_preds, emoji_dev_preds, emoji_test_preds)
def run_sent140_emojiless(self):
'''
Trained on sent140, fine-tuned on emojiless, predict on emoji
Report score on sent 140 too, just because it's intersting
'''
# model
self.model_sent140_emojiless = TorchShallowNeuralClassifier()
# train
# combine features
combined_train_X = np.vstack((self.sent140_train_X, self.emojiless_train_X))
combined_train_Y = self.sent140_train_Y + self.emojiless_train_Y
self.model_sent140_emojiless.fit(combined_train_X, combined_train_Y)
# test on sent140
sent140_train_preds = self.model_sent140_emojiless.predict(self.sent140_train_X)
sent140_dev_preds = self.model_sent140_emojiless.predict(self.sent140_dev_X)
# test on emoji
emoji_train_preds = self.model_sent140_emojiless.predict(self.emoji_train_X)
emoji_dev_preds = self.model_sent140_emojiless.predict(self.emoji_dev_X)
if self.testing:
emoji_test_preds = self.model_sent140_emojiless.predict(self.emoji_test_X)
else:
emoji_test_preds = None
return (sent140_train_preds, sent140_dev_preds, emoji_train_preds, emoji_dev_preds, emoji_test_preds)
def run_sent140_emoji(self):
'''
Trained on sent140, fine-tuned on emoji, predict on emoji
Report score on sent 140 too, just because it's intersting
'''
# model
self.model_sent140_emoji = TorchShallowNeuralClassifier()
# train
combined_train_X = np.vstack((self.sent140_train_X, self.emoji_train_X))
combined_train_Y = self.sent140_train_Y + self.emoji_train_Y
self.model_sent140_emoji.fit(combined_train_X, combined_train_Y)
# test on sent140
sent140_train_preds = self.model_sent140_emoji.predict(self.sent140_train_X)
sent140_dev_preds = self.model_sent140_emoji.predict(self.sent140_dev_X)
# test on emoji
emoji_train_preds = self.model_sent140_emoji.predict(self.emoji_train_X)
emoji_dev_preds = self.model_sent140_emoji.predict(self.emoji_dev_X)
if self.testing:
emoji_test_preds = self.model_sent140_emoji.predict(self.emoji_test_X)
else:
emoji_test_preds = None
return (sent140_train_preds, sent140_dev_preds, emoji_train_preds, emoji_dev_preds, emoji_test_preds)