Searched refs:tensor_tile_rank (Results 1 – 1 of 1) sorted by relevance
606 int64 tensor_tile_rank = sharding.tile_assignment_dimensions_size(); in GetDimensionIndicesAndNumSplitsFromSharding() local608 tensor_tile_rank--; in GetDimensionIndicesAndNumSplitsFromSharding()610 for (int dim_index = 0; dim_index < tensor_tile_rank; dim_index++) { in GetDimensionIndicesAndNumSplitsFromSharding()