Directly randomly sample a batch of data in, and we do not manually do anything specific to the long tail data ?

jiequancui / ResLT

ResLT: Residual Learning for Long-tailed Recognition (TPAMI 2022)

https://arxiv.org/pdf/2101.10633.pdf

MIT License

56 stars 5 forks source link

Closed Zhaoyi-Yan closed 2 years ago

Zhaoyi-Yan commented 2 years ago

Am I correct ?

jiequancui commented 2 years ago

Hi,

We randomly sample a batch of data from the dataset for each iteration training.