tfswin 1.1.0

Keras (TensorFlow v2) reimplementation of Swin Transformer model.

tfswin

Keras (TensorFlow v2) reimplementation of Swin Transformer model.

Based on Official Pytorch implementation.

Examples

Default usage:

from tfswin import SwinTransformerTiny224  # + 5 other variants

model = SwinTransformerTiny224()  # by default will download imagenet[21k]-pretrained weights
model.compile(...)
model.fit(...)

Custom classification:

from keras import layers, models
from tfswin import SwinTransformerTiny224

base_model = SwinTransformerTiny224(include_top=False)
new_outputs = layers.Dense(100, activation='softmax')(base_model.outputs)
new_model = models.Model(inputs=base_model.inputs, outputs=new_outputs)

new_model.compile(...)
new_model.fit(...)

Differences

Code simplification:

• Input height and width are always equal
• Patch height and width are always equal
• All input shapes automatically evaluated (not passed through a constructor)

Performance improvements:

• Layer normalization epsilon fixed at 1.001e-5, inputs are casted to float32 to use fused op implementation.
• Some layers (like PatchMerging) have been refactored to use faster TF operations.

Citation

@article{liu2021Swin,
  title={Swin Transformer: Hierarchical Vision Transformer using Shifted Windows},
  author={Liu, Ze and Lin, Yutong and Cao, Yue and Hu, Han and Wei, Yixuan and Zhang, Zheng and Lin, Stephen and Guo, Baining},
  journal={arXiv preprint arXiv:2103.14030},
  year={2021}
}