Classifier Models

Community repository of pre-trained models for the classifier gem.

Using Models

# List available models
classifier models

# Download a model
classifier pull spam-filter

# Use a model directly (downloads if needed)
classifier -r spam-filter "Is this spam?"

# Use from a different registry
classifier -r @other-user/models:sentiment "I love this!"

Available Models

Name	Type	Description
Coming soon

Contributing a Model

1. Train your classifier:

classifier train spam spam_corpus.txt
classifier train ham ham_corpus.txt

2. Submit via CLI:

classifier push ./classifier.json --name "my-model" --description "Description here"

Or manually:

1. Fork this repository
2. Add your model JSON to models/
3. Update models.json with metadata
4. Submit a pull request

Model JSON Format

Models are standard classifier JSON exports:

{
  "type": "bayes",
  "categories": ["spam", "ham"],
  "...classifier data..."
}

models.json Entry

{
  "my-model": {
    "description": "What this model classifies",
    "type": "bayes",
    "categories": ["cat1", "cat2"],
    "file": "models/my-model.json",
    "version": "1.0.0",
    "author": "github-username"
  }
}

Creating Your Own Registry

Any GitHub repo with this structure works as a registry:

your-repo/
├── models/
│   └── your-model.json
└── models.json

Users can then use your models:

classifier -r @your-username/your-repo:your-model "text"
classifier pull @your-username/your-repo:your-model
classifier pull @your-username/your-repo  # all models

Guidelines

• Models should be generally useful (not project-specific)
• Include clear descriptions
• Test before submitting
• No sensitive/private training data
• Reasonable file sizes (< 10MB preferred)

License

Models contributed here are licensed under MIT unless otherwise specified by the author.