Code implementation in PyTorch of the paper "The Hedgehog & the Porcupine: Expressive Linear Attentions with Softmax Mimicry"
Paper Link https://arxiv.org/abs/2402.04347
- Lucidrains
- Agora Research
Try running example.py
Apache 2.0
@misc{zhang2024hedgehog, title={The Hedgehog & the Porcupine: Expressive Linear Attentions with Softmax Mimicry}, author={Michael Zhang and Kush Bhatia and Hermann Kumbong and Christopher RĂ©}, year={2024}, eprint={2402.04347}, archivePrefix={arXiv}, primaryClass={cs.LG} }