This is the initial release of Nystromformer and implements Nystromformer as a subclassed TensorFlow model.
- [MoorePenrosePseudoinverse](https://github.com/Rishit-dagli/Nystromformer/blob/main/nystromformer/utils.py#L5-L39): Implement the Moore Penrose Inverse as a Keras layer
- [NystromAttention](https://github.com/Rishit-dagli/Nystromformer/blob/main/nystromformer/nystrom_attention.py#L9-L164): Implment Nystrom Attention as a Keras layer
- [Nystromformer](https://github.com/Rishit-dagli/Nystromformer/blob/main/nystromformer/nystromformer.py#L7-L49): Put together a complete Nystromformer model stacking Nystrom Attention.
- [PreNorm](https://github.com/Rishit-dagli/Nystromformer/blob/main/nystromformer/utils.py#L42-L50): Normalize the activations of the previous layer for each given example in a batch independently and apply some function to it, implemented as a TensorFlow Keras Layer.
- [FeedForward](https://github.com/Rishit-dagli/Nystromformer/blob/main/nystromformer/utils.py#L53-L67): Create a FeedForward neural net with two Dense layers and GELU activation, implemented as a TensorFlow Keras Layer.
**Full Changelog**: https://github.com/Rishit-dagli/Nystromformer/commits/v0.1.0