stanfordnlp/CoreNLP · error · RuntimeIOException

Error loading classifier from

Error message

Error loading classifier from 

What it means

MultinomialLogisticClassifier.loadSelfSupervised/load deserializes a classifier (weights array, feature index, label index) from a file via ObjectInputStream. If the stream is corrupt, truncated, or not a saved classifier, an IOException or ClassNotFoundException is wrapped in this RuntimeIOException carrying the path. It is a deserialization failure, not a modeling error.

Solutions

1. Verify the path points to a file previously written by MultinomialLogisticClassifier.save with the same library version
2. Catch RuntimeIOException and check the cause (IOException vs ClassNotFoundException) to distinguish corrupt file from version mismatch
3. Re-save the classifier with the current library version and retry
4. Check file existence/readability before load

Example fix

// before
MultinomialLogisticClassifier c = MultinomialLogisticClassifier.load(cfgPath);
// after
File f = new File(modelPath);
if (!f.isFile()) throw new IllegalArgumentException("model missing: " + modelPath);
try {
  MultinomialLogisticClassifier c = MultinomialLogisticClassifier.load(modelPath);
} catch (RuntimeIOException e) {
  throw new IllegalStateException("Incompatible/corrupt model at " + modelPath, e);
}

File f = new File(path);
if (!f.exists() || !f.canRead() || f.length() < 8)
  throw new IllegalArgumentException("Classifier file missing or empty: " + path);

try {
  MultinomialLogisticClassifier<LL,FF> c = MultinomialLogisticClassifier.load(path);
} catch (RuntimeIOException e) {
  if (e.getCause() instanceof ClassNotFoundException)
    throw new IllegalStateException("Model serialized with a different library version: " + path, e);
  throw new IllegalStateException("Corrupt or missing model file: " + path, e);
}

When it happens

Trigger: Calling MultinomialLogisticClassifier.load(path) on a missing, corrupt, or wrong-format file; loading a classifier serialized by an incompatible library version (class shape changed → ClassNotFoundException); reading a text file or classifier of a different type.

Common situations: Pointing to the wrong path or a file moved/truncated; version mismatch after upgrading CoreNLP so the serialized class descriptor differs; attempting to load a LinearClassifier's dump as a MultinomialLogisticClassifier.

Understand the failure class

Background: "failed to read file", EACCES, ENOENT and "could not read <path>" errors: when a program can't read a file from disk — this error's family across 49 libraries.

Related errors

• Error leading weights from
• Couldn't load
• IO problem reading classifier.
• Failed to load segmenter
• Failed to read lambdas from given input stream

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/a188eb5a674ce207. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/classify/MultinomialLogisticClassifier.java:126

  }

  @Override
  public Counter<L> logProbabilityOf(Datum<L, F> example) {
    Counter<L> result = probabilityOf(example);
    Counters.logInPlace(result);
    return result;
  }

  private static <LL,FF> MultinomialLogisticClassifier<LL,FF> load(String path) {
    Timing t = new Timing();
    try (ObjectInputStream in = IOUtils.readStreamFromString(path)) {
      double[][] myWeights = ErasureUtils.uncheckedCast(in.readObject());
      Index<FF> myFeatureIndex = ErasureUtils.uncheckedCast(in.readObject());
      Index<LL> myLabelIndex = ErasureUtils.uncheckedCast(in.readObject());
      t.done(logger, "Loading classifier from " + path);
      return new MultinomialLogisticClassifier<>(myWeights, myFeatureIndex, myLabelIndex);
    } catch (IOException | ClassNotFoundException e) {
      throw new RuntimeIOException("Error loading classifier from " + path, e);
    }
  }

  private void save(String path) throws IOException {
    System.out.print("Saving classifier to " + path + "... ");

    // make sure the directory specified by path exists
    int lastSlash = path.lastIndexOf(File.separator);
    if (lastSlash > 0) {
      File dir = new File(path.substring(0, lastSlash));
      if (! dir.exists())
        dir.mkdirs();
    }

    ObjectOutputStream out = new ObjectOutputStream(new FileOutputStream(path));
    out.writeObject(weights);
    out.writeObject(featureIndex);
    out.writeObject(labelIndex);

View on GitHub (pinned to 1b7edd19c4)