# script for corresponding test case
# most test cases should be able to be executed without any further changes, if data is available
SAMPLE_SIZE = 100000
EPOCHS = 150
HIDDEN_SIZE = 10000
BATCH_SIZE = 128
# change device to "cpu" if cuda not available
DEVICE = "cuda"
stopwatch = StopWatch()
# pregenerated embedding and labels
file_words = open("../tests/embedding8.json")
file_labels = open("../data/unique_labels.json")
data_provider = DataProviderLight(file_words,
file_labels,
sample_size=SAMPLE_SIZE)
# embedding data, splitting up into train and test set
processor = PregeneratedProcessor(data_provider)
generator = DeterministicGenerator(data_provider, processor)
dataset = generator.generate_dataset()
# creating classifier, overwriting parameters
classifier = FFNClassifier(data_provider, dataset, DEVICE, BATCH_SIZE,
HIDDEN_SIZE)
# train classifier and output progress
trainer = ClassificationTrainer(classifier)
trainer.enable_loss(10, True)
trainer.enable_precision(10, True)
trainer.enable_trainset_precision(10, True)
trainer.train(EPOCHS)
trainer.show()
print("done")
# change device to "cpu" if cuda not available
DEVICE = "cuda"
stopwatch = StopWatch()
# pregenerated embedding and labels
file_words = open("../tests/embedding1.json")
file_labels = open("../data/unique_labels.json")
data_provider = DataProviderLight(file_words,
file_labels,
sample_size=SAMPLE_SIZE)
# embedding data, splitting up into train and test set
processor = PregeneratedProcessor(data_provider)
generator = DeterministicGenerator(data_provider, processor)
dataset = generator.generate_dataset()
for hidden_size in HIDDEN_SIZE:
print("hidden size: " + str(hidden_size))
# creating classifier, overwriting parameters
classifier = FFNClassifier(data_provider,
dataset,
DEVICE,
batch_size=BATCH_SIZE,
hidden_size=hidden_size)
# train classifier and output progress
trainer = ClassificationTrainer(classifier)
trainer.enable_loss(10, True)
stopwatch.start()
trainer.train(EPOCHS)
stopwatch.stop()
# test not really necessary, done by trainer
test = ClassificationTest(dataset, classifier)
print("Präzision: " + str(test.test()) + "%")
print("Präzision (Trainingsset): " + str(test.test_trainset()) + "%")