0.1.3

Major Features and Improvements

• Add type hints for all public APIs
• Replace *.available_feature_keys() to *.available_feature_keys (using static variable)
• Add *.available_weights

What's Changed

• Add Colab examples by @james77777778 in #24
• Add type hints by @james77777778 in #25
• Improve test coverage and refactor BaseModel by @james77777778 in #26

Full Changelog: 0.1.2...0.1.3

0.1.2

Major Features and Improvements

Now, almost all models default to wegiths="imagenet".

What's Changed

• Add pypi support by @james77777778 in #17
• Defaults to weights=imagenet if possible by @james77777778 in #19
• Fix InceptionV3 and add Codecov by @james77777778 in #20

Full Changelog: 0.1.0...0.1.2

First Release of `kimm`

Description

Please see README for the details.

Currently Supported Models with Pretrained Weights

Classics

• DenseNet
• InceptionV3
• ResNet
• VGG
• Xception

Lightweight

• EfficientNetLite
• GhostNet
• GhostNetV2
• LCNet
• MobileNetV2
• MobileNetV3
• MobileViT
• TinyNet

SOTA

• ConvMixer
• ConvNeXt
• EfficientNet
• EfficientNetV2
• RegNet
• ViT

New Contributors

• @james77777778 made their first contribution in #3

Full Changelog: https://github.com/james77777778/kimm/commits/0.1.0