About common/rela.py & routing_transformer.py & linearattention.py

JinluZhang1126 / MixSTE

Official implementation of CVPR 2022 paper(MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Video)

193 stars 17 forks source link

About common/rela.py & routing_transformer.py & linearattention.py #14

Closed Z-Z-J closed 2 years ago

Z-Z-J commented 2 years ago

QQ截图20221018162655 Can you provide these files ?

JinluZhang1126 commented 2 years ago

Can you provide these files ?

These files are conducted from other works, which mainly focus on decreasing the computation cost of Transformer. They do not affect the results of the model. We only use them to test the effects of our work.

If you have interest in it, please pay attention to these works:

Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting;
Reformer: The Efficient Transformer
Linformer: Self-Attention with Linear Complexity