[Kaggle Model] Video Swin Transformer in TF 2, Keras V3, TFLite, ONNX

Show and Tell
, , ,
#1

Hello. We have published Video Swin Transformer in Kaggle Model hub. The checkpoints are available in TensorFlow (Keras V2), Keras V3, TensorFlow Lite and ONNX format. Check out the following resources.

Kaggle model:

GitHub repo:

Highlights

  1. TensorFlow (Keras V2)

carbon
carbon2132×1636 361 KB

  1. Keras V3

carbon (1)
carbon (1)2468×2084 483 KB

  1. TensorFlow Lite

carbon (2)
carbon (2)1920×1731 146 KB

  1. ONNX

carbon (3)
carbon (3)1920×1251 94.9 KB