Keras

Keras is a high-level neural network API written in Python, originally created by François Chollet and merged into TensorFlow in 2017, but still usable as a standalone framework. It is an open-source deep learning framework that runs on top of deep learning backends such as TensorFlow, Theano, or Microsoft Cognitive Toolkit (CNTK).

You can use Keras to quickly train models while using SwanLab for experiment tracking and visualization.

Online Demo

1. Import SwanLabLogger

from swanlab.integration.keras import SwanLabLogger

2. Integrate with model.fit

First, initialize SwanLab:

swanlab.init(
    project="keras_mnist",
    experiment_name="mnist_example",
    description="Keras MNIST Example"
    )

Then, add SwanLabLogger to the callbacks parameter in model.fit to complete the integration:

model.fit(..., callbacks=[SwanLabLogger()])

3. Example - MNIST

from swanlab.integration.keras import SwanLabLogger
import tensorflow as tf
import swanlab

# Initialize SwanLab
swanlab.init(
    project="keras_mnist",
    experiment_name="mnist_example",
    description="Keras MNIST Example"
    )

# Load and preprocess MNIST data
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()
x_train = x_train.reshape(-1, 28, 28, 1).astype('float32') / 255.0
x_test = x_test.reshape(-1, 28, 28, 1).astype('float32') / 255.0

# Build a simple CNN model
model = tf.keras.Sequential([
    tf.keras.layers.Conv2D(32, 3, activation='relu', input_shape=(28, 28, 1)),
    tf.keras.layers.MaxPooling2D(),
    tf.keras.layers.Conv2D(64, 3, activation='relu'),
    tf.keras.layers.MaxPooling2D(),
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.Dense(10, activation='softmax')
])

# Compile the model
model.compile(
    optimizer='adam',
    loss='sparse_categorical_crossentropy',
    metrics=['accuracy']
)

# Train the model with SwanLabLogger
model.fit(
    x_train, 
    y_train,
    epochs=5,
    validation_data=(x_test, y_test),
    callbacks=[SwanLabLogger()]
)

Effect demonstration:

Online Demo