Release v1.4
Upgraded required scipy version and graceful handling of NaNs in objective evaluation in trust-region step. No change to expected behavior of DFO-LS algorithm.
Upgraded required scipy version and graceful handling of NaNs in objective evaluation in trust-region step. No change to expected behavior of DFO-LS algorithm.