keystoneml.nodes.learning.internal
Use BCD to solve W = (X.
Use BCD to solve W = (X.t * (diag(B) * X) + \lambda * I) \ X.t * (B .* Y)
blockSize to use for Block Coordinate Descent
number of blocks in the input
number of iterations of BCD to run
L2 regularization parameter
number of features in the input (columns in X)
number of classes (columns in Y)
blocks of input features
zero-mean'd labels matrix
diagonal values for the weights matrix
mean to be subtracted from X before the solve