def eval_seizure_model(model, model_type):
example = io.StringIO(example_seizure_file)
dataset = load_seizure_dataset(example, model_type)
loader = torch.utils.data.DataLoader(dataset,
batch_size=3,
shuffle=False,
num_workers=0)
return model_eval(model, loader)
def test_load_seizure_dataset_mlp():
example = io.StringIO(example_seizure_file)
tensor_dataset = load_seizure_dataset(example, 'MLP')
expect(
type(tensor_dataset) == torch.utils.data.dataset.TensorDataset,
"it should return TensorDataset")
data_tensor, target_tensor = tensor_dataset.tensors
expect(data_tensor.size() == (3, 178), "it does not have expected shapes")
expect(target_tensor.size() == (3, ), "it does not have expected shapes")
assert_expectations()
PATH_TRAIN_FILE = "../data/seizure/seizure_train.csv"
PATH_VALID_FILE = "../data/seizure/seizure_validation.csv"
PATH_TEST_FILE = "../data/seizure/seizure_test.csv"
# Path for saving model
PATH_OUTPUT = "../output/seizure/"
os.makedirs(PATH_OUTPUT, exist_ok=True)
# Some parameters
MODEL_TYPE = 'MLP' # TODO: Change this to 'MLP', 'CNN', or 'RNN' according to your task
NUM_EPOCHS = 3
BATCH_SIZE = 32
USE_CUDA = False # Set 'True' if you want to use GPU
NUM_WORKERS = 0 # Number of threads used by DataLoader. You can adjust this according to your machine spec.
train_dataset = load_seizure_dataset(PATH_TRAIN_FILE, MODEL_TYPE)
valid_dataset = load_seizure_dataset(PATH_VALID_FILE, MODEL_TYPE)
test_dataset = load_seizure_dataset(PATH_TEST_FILE, MODEL_TYPE)
train_loader = torch.utils.data.DataLoader(train_dataset,
batch_size=BATCH_SIZE,
shuffle=True,
num_workers=NUM_WORKERS)
valid_loader = torch.utils.data.DataLoader(valid_dataset,
batch_size=BATCH_SIZE,
shuffle=False,
num_workers=NUM_WORKERS)
test_loader = torch.utils.data.DataLoader(test_dataset,
batch_size=BATCH_SIZE,
shuffle=False,
num_workers=NUM_WORKERS)
