jolars/sgdnet

Issues

• Faster SubSampling for Minibatches

  #30 opened 5 years ago by michaelweylandt
  3

• Warn if maximum number of iterations was met.

  #25 opened 5 years ago by jolars
  0

• Fix unwarranted destruction of sparsity while calculating sample norms

  #23 opened 5 years ago by jolars
  0

• First fit on automatic lambda path for ridge penalty

  #19 opened 6 years ago by jolars
  4

• Erroneous results with non-standardized features

  #12 opened 6 years ago by jolars
  8

• Data in package and for benchmarking

  #13 opened 6 years ago by jolars
  4

• Approximation of lipschitz constant for use in step size

  #7 opened 6 years ago by jolars
  3

• Save a data copy in fitted objects

  #10 opened 6 years ago by jolars
  0

• Separating sparse and dense implementations

  #6 opened 6 years ago by jolars
  2

• lambda_{max} for multivariate gaussian model

  #3 opened 7 years ago by jolars
  5

• Fitting intercept differs between sparse and dense implementations

  #2 opened 7 years ago by jolars
  0