Native-sparse-attention-pytorch

**Full Changelog**: https://github.com/lucidrains/native-sparse-attention-pytorch/compare/0.1.21...01.22

What's Changed
* compress_block_sliding_stride by Mr-Grin in https://github.com/lucidrains/native-sparse-attention-pytorch/pull/25


**Full Changelog**: https://github.com/lucidrains/native-sparse-attention-pytorch/compare/0.1.27...0.2.0

What's Changed
* Small change so token embeddings aren't looked up for past tokens during inference by Pasewark in https://github.com/lucidrains/native-sparse-attention-pytorch/pull/24


**Full Changelog**: https://github.com/lucidrains/native-sparse-attention-pytorch/compare/0.1.26...0.1.27

**Full Changelog**: https://github.com/lucidrains/native-sparse-attention-pytorch/compare/0.1.25...0.1.26

None

What's Changed
* Added CompressTransformer, a simple transformer to compress the blocks by Pasewark in https://github.com/lucidrains/native-sparse-attention-pytorch/pull/21

New Contributors
* Pasewark made their first contribution in https://github.com/lucidrains/native-sparse-attention-pytorch/pull/21

**Full Changelog**: https://github.com/lucidrains/native-sparse-attention-pytorch/compare/0.1.23...0.1.24