Open PlutoQyl opened 1 year ago
I noticed that the batchsize dimension is concatenated together when selecting the top 1% learning region, does this cause the batchsize to have a big impact on performance?
I noticed that the batchsize dimension is concatenated together when selecting the top 1% learning region, does this cause the batchsize to have a big impact on performance?