Starring is not enough and I need to keep track of why repo has interest. So here is a list of TensorFlow 2.x projects on Github that sparked my interest for some reason.
| Github Project | Why |
|---|---|
| PyCon SG 2019: Optimize TF perf | Setup XLA, MPT on ResNet and Transformers |
| TensorFlow Profiler | Detailled profiling library in TensorBoard |
| Github Project | Why |
|---|---|
| Super Resolution | Clean implementation of multiple SR models with tf.GradientTape |
| YOLOv3 | YOLOv3 implementation with nice definition of backbone, head and loss |
| BERT | tf.keras.Layer implementation of BERT (embeddings, attention) |
| Keras-FewShotLearning | Playground for few shot learning classification |
| Github Project | Why |
|---|---|
| TensorFlow Determinism | Determinism and non-determinism at GPU level (from NVidia) |
| tf-explain | Intepretability of DL models for TF2.x |
| Github Project | Why |
|---|---|
| TensorFlow-Examples | Recipes for TF2.x |