Examples of OrthantWiseLimitedMemoryBFGS

cc.mallet.optimize.OrthantWiseLimitedMemoryBFGS
Implementation of orthant-wise limited memory quasi Newton method for optimizing convex L1-regularized objectives. See: "Scalable training of l1-regularized log-linear models" by Galen Andrew and Jianfeng Gao in ICML 2007 for details. This code is an adaptation of the freely-available C++ code on Galen's webpage. @author Kedar Bellare

Examples of cc.mallet.optimize.OrthantWiseLimitedMemoryBFGS

    this.gaussianPriorVariance = Double.MAX_VALUE;
  }


  public Optimizer getOptimizer() {
    if (optimizer == null && optimizable != null)
      optimizer = new OrthantWiseLimitedMemoryBFGS(optimizable, l1Weight);
    return optimizer;
  }

View Full Code Here

Examples of cc.mallet.optimize.OrthantWiseLimitedMemoryBFGS

    if (trainingSet != this.trainingSet || optimizable == null) {
      getOptimizable(trainingSet);
      optimizer = null;
    }
    if (optimizer == null)
      optimizer = new OrthantWiseLimitedMemoryBFGS(optimizable, l1Weight);
    return optimizer;
  }

View Full Code Here

Examples of cc.mallet.optimize.OrthantWiseLimitedMemoryBFGS

  }


  public Optimizer getOptimizer(InstanceList trainingSet) {
    getOptimizableCRF(trainingSet);
    if (opt == null || ocrf != opt.getOptimizable())
      opt = new OrthantWiseLimitedMemoryBFGS(ocrf, l1Weight);
    return opt;
  }

View Full Code Here

TOP

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.