flowa - Machine Learning Toolkit

flowa: (V10.5.5)

Python Machine Learning, Image Generation, Decision Trees, Label Encoders, Sequential, and more!

Installing

# Linux/macOS
python3 pip install -U flowa

# Windows
py -3 -m pip install -U flowa

Simple Examples

x = flowa.Array([[0, 0], [0, 1], [1, 0], [1, 1]])
y = flowa.Array([[0], [1], [1], [0]])

network = flowa.Network(
    flowa.Input(2),
    (
        flowa.Hidden(4, flowa.Tanh), 
        flowa.Hidden(2, flowa.Sigmoid)
    ),
    flowa.Output(1)
)

network.train(x, y, epoch=1000)
print(network.predict(x))

from flowa.ai import (
    Encoder,
    Tree,
    Dataset,
    read_csv,
    convert
)

classifier: Tree = Tree()
encoder: Encoder = Encoder()

dataset: str = convert(Dataset.get_music_data())
csv: object = read_csv(dataset)

dataframe: object = encoder.df(csv, 'gender')

X_matrix: object = dataframe.drop('genre', axis=1).values
y_column: object = encoder(dataframe['genre'].values)

classifier.fit(X_matrix, y_column)

age, gender = encoder.new(30, 'female')
fix: list = encoder.fix(age, gender)

prediction: list[int] = classifier.predict(fix)
print(encoder.inverse(prediction))

#>>> ['Pop']

Image generation:

model: ImageModel[object] = ImageModel()
image: ImageModel[str] = model.generate(
    prompt="a cat", model="pixart", width=512, height=512
).save("some-file.png")

#>>> flowa.types.Image

String Dataset to dataframe conversion:

from flowa.ai import (
    Dataset,
    read_csv,
    convert
)

dataset: Dataset = Dataset.get_play_tennis()

converted_dataset: str = convert(dataset)

csv: Dataset = read_csv(converted_dataset)
print(csv)

#>>>     Outlook Temperature Humidity    Wind Play Tennis
#>>> 0  Overcast        Mild   Normal    Weak         Yes
#>>> 1     Sunny        Mild   Normal    Weak         Yes
#>>> ... [2 rows not shown]

Links

• Github
• Project