What causes "Unsupported minimizer ${MINIMIZER}" in stanfordnlp/CoreNLP?

Set MINIMIZER to a value handled by the switch (e.g. the AdaGrad/SGD constant used in the release)

stanfordnlp/CoreNLP · error · java.lang.IllegalArgumentException

Unsupported minimizer

Error message

Unsupported minimizer ${MINIMIZER}

What it means

DVParser.executeOneTrainingBatch switches on the MINIMIZER constant to run the selected optimizer (e.g. SGD/AdaGrad variants); the default branch throws IllegalArgumentException for unsupported minimizer values. This is an internal configuration guard — MINIMIZER is a class-level constant not normally exposed as a CLI flag.

Solutions

Set MINIMIZER to a value handled by the switch (e.g. the AdaGrad/SGD constant used in the release)
Remove custom minimizer overrides and use the stock DVParser build
If adding a new optimizer, extend the switch in executeOneTrainingBatch before training

Example fix

// before
private static final int MINIMIZER = 42; // unsupported
// after
private static final int MINIMIZER = DVParser.ADA_GRAD;

Defensive patterns

Strategy: validation

Validate before calling

// MINIMIZER is a source-level constant in DVParser
if (MINIMIZER != DVParser.ADA_GRAD /* and other supported constants */) {
    throw new IllegalArgumentException("MINIMIZER value not supported by executeOneTrainingBatch");
}

Try / catch

try {
    parser.train(trainTreebank, compTreebank);
} catch (IllegalArgumentException e) {
    if (e.getMessage().contains("Unsupported minimizer")) {
        System.err.println("Reset MINIMIZER to a value handled by the training switch");
    }
}

Prevention

Never change the MINIMIZER constant without updating the switch in executeOneTrainingBatch
Use stock DVParser builds for training
Cover new optimizer values with tests that run one training batch

When it happens

Trigger: A developer editing DVParser source sets MINIMIZER to a value not covered by the switch cases, then calls train, which invokes executeOneTrainingBatch during a training iteration.

Common situations: Custom builds of CoreNLP where someone added or changed the minimizer constant; code that programmatically modifies DVParser configuration; version drift where a chosen minimizer name no longer matches the switch.

Understand the failure class

Background: "Invalid value" and "allowed values are" config errors: what your library rejected and how to fix it — this error's family across 41 libraries.

Related errors

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/d1997f7e143c8b84. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/parser/dvparser/DVParser.java:296

      break;
    }
    case 3:{
      // AdaGrad
      double eps = 1e-3;
      double currCost = 0;
      for(int i = 0; i < op.trainOptions.qnIterationsPerBatch; i++){
        double[] gradf = gcFunc.derivativeAt(theta);
        currCost = gcFunc.valueAt(theta);
        log.info("batch cost: " + currCost);
        for (int feature =0; feature<gradf.length;feature++ ) {
          sumGradSquare[feature] = sumGradSquare[feature] + gradf[feature]*gradf[feature];
          theta[feature] = theta[feature] - (op.trainOptions.learningRate * gradf[feature]/(Math.sqrt(sumGradSquare[feature])+eps));
        }
      }
      break;
    }
    default: {
      throw new IllegalArgumentException("Unsupported minimizer " + MINIMIZER);
    }
    }


    dvModel.vectorToParams(theta);
  }

  public DVParser(DVModel model, LexicalizedParser parser) {
    this.parser = parser;
    this.op = parser.getOp();
    this.dvModel = model;
  }

  public DVParser(LexicalizedParser parser) {
    this.parser = parser;
    this.op = parser.getOp();

    if (op.trainOptions.randomSeed == 0) {

View on GitHub (pinned to 1b7edd19c4)