Our customized keras package for popular image classifier techniques.
Project description
wyn-keras 🎉
A Python package for building and experimenting with Vision Transformer (ViT) models using TensorFlow and Keras.
Directory Structure 📁
wyn-keras/
├── pyproject.toml
├── README.md
├── wyn_keras
│ ├── __init__.py
│ └── vit.py
├── tests
│ └── __init__.py
└── .gitignore
Installation Instructions 📦
To install the package and its dependencies, use Poetry:
# Install Poetry if you haven't already
curl -sSL https://install.python-poetry.org | python3 -
# Install the package
poetry install
Usage 🚀
Vision Transformer
The ViT
class allows you to create and train Vision Transformer models.
Additional Functions (Coming Soon...) 🚧
Stay tuned for more functionalities to be added in the future!
Example Usage 📚
MNIST Example
import tensorflow as tf
from wyn_keras.vit import ViT
# Load and preprocess the MNIST dataset
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()
x_train = x_train[..., tf.newaxis].astype("float32") / 255.0
x_test = x_test[..., tf.newaxis].astype("float32") / 255.0
# Number of classes in MNIST dataset
num_classes = 10
# Create an instance of the ViT class
vit_model = ViT(num_classes=num_classes, input_shape=(28, 28, 1), image_size=28, num_epochs=2)
# Create the ViT model
model = vit_model.create_vit_classifier()
# Train the model
history = vit_model.run_experiment(model, x_train, y_train, x_test, y_test)
# Plot attention
vit_model.plot_attention(model, x_test[0])
CIFAR-10 Example
import tensorflow as tf
from wyn_keras.vit import ViT
# Load and preprocess the CIFAR-10 dataset
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.cifar10.load_data()
x_train = x_train.astype("float32") / 255.0
x_test = x_test.astype("float32") / 255.0
# Number of classes in CIFAR-10 dataset
num_classes = 10
# Create an instance of the ViT class
vit_model = ViT(num_classes=num_classes, input_shape=(32, 32, 3), image_size=32, num_epochs=2)
# Create the ViT model
model = vit_model.create_vit_classifier()
# Train the model
history = vit_model.run_experiment(model, x_train, y_train, x_test, y_test)
# Plot attention
vit_model.plot_attention(model, x_test[0])
Author ✍️
Yiqiao Yin
Email: eagle0504@gmail.com
Personal Site: https://www.y-yin.io/
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
wyn_keras-0.1.3.tar.gz
(4.9 kB
view hashes)
Built Distribution
Close
Hashes for wyn_keras-0.1.3-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 168bd558089d3ee75fa98195e76b9dea3cb08403043c259711a6a05e36d94972 |
|
MD5 | 21923f336c292acdfb57c41461eeee68 |
|
BLAKE2b-256 | b72e92130bad9754dc069dbfbec5bc68acd0abe10751e31fb9d2d8f7eaabec87 |