[GitHub] eric-haibin-lin commented on issue #11299: row_sparse_pull, push row_sparse gradient is too slow, it has 10+ times difference

eric-haibin-lin commented on issue #11299: row_sparse_pull,push row_sparse gradient is too slow,it has 10+ times difference URL: https://github.com/apache/incubator-mxnet/issues/11299#issuecomment-398503457 @coldsheephot are you working on multi-device or multi-machine case? I have plan

[GitHub] eric-haibin-lin commented on issue #11299: row_sparse_pull, push row_sparse gradient is too slow, it has 10+ times difference

eric-haibin-lin commented on issue #11299: row_sparse_pull,push row_sparse gradient is too slow,it has 10+ times difference URL: https://github.com/apache/incubator-mxnet/issues/11299#issuecomment-397846747 For minibatch training usually you derive rowids from the sparse data in the

[GitHub] eric-haibin-lin commented on issue #11299: row_sparse_pull, push row_sparse gradient is too slow, it has 10+ times difference

eric-haibin-lin commented on issue #11299: row_sparse_pull,push row_sparse gradient is too slow,it has 10+ times difference URL: https://github.com/apache/incubator-mxnet/issues/11299#issuecomment-397521477 row_sparse_pull will be fast only when the number of row ids is very small e.g.

[GitHub] eric-haibin-lin commented on issue #11299: row_sparse_pull, push row_sparse gradient is too slow, it has 10+ times difference

eric-haibin-lin commented on issue #11299: row_sparse_pull,push row_sparse gradient is too slow,it has 10+ times difference URL: https://github.com/apache/incubator-mxnet/issues/11299#issuecomment-397521477 row_sparse_pull will be fast only when the number of row ids is very small