Spell Classifier

Neural network that classifies spell commands into 4 categories.

Architecture: Word2Vec embeddings (50-dim) → MLP classifier (32→16 hidden)

Categories: offensive, defensive, healing, utility

Usage

poetry install

# Generate training data (8000 samples)
poetry run python -m spell_neural.generate_data

# Train model
poetry run python -m spell_neural.train_model

# Interactive prediction
poetry run python -m spell_neural.predictor