Enhancing the Locality and Breaking the Memory Bottleneck of Transformer on Time Series Forecasting

Benchmark Model Rank Results
image-generation-on-imagenet-64x64Logsparse (6 layers)#51Bits per dim: 4.351