def setup(self, stage: str = None):
df = pd.read_csv(f"{pathlib.Path(__file__).parent.absolute()}/{DATA_FILENAME}")
# Split features, targets
x = df.drop(columns=[Y_LABEL]).values
y = df[Y_LABEL].values
UCIDataModule.setup(self, x, y)
def setup(self, stage: str = None):
df = pd.read_csv(
f"{pathlib.Path(__file__).parent.absolute()}/{DATA_FILENAME}",
sep=";")
x = df.drop(columns=Y_LABEL).values
y = df[Y_LABEL].values
UCIDataModule.setup(self, x, y)
def setup(self, stage: str = None):
df = pd.read_csv(f"{pathlib.Path(__file__).parent.absolute()}/{DATA_FILENAME}")
# Loads rows as string
data = np.empty((len(df.index), self.dims + self.out_dims))
for i in range(data.shape[0]):
data[i] = np.array(
[float(el) for el in df.values[i][0].split(" ") if el != ""]
)
x = data[:, :-1]
y = data[:, -1]
UCIDataModule.setup(self, x, y)
def __init__(
self,
batch_size: int,
n_workers: int,
train_val_split: float = 0.9,
test_split: float = 0.1,
**kwargs,
):
UCIDataModule.__init__(
self,
batch_size,
n_workers,
train_val_split,
test_split,
)
# Manual as we know it
self.dims = 4
self.out_dims = 1