Skip to content

Getting Started

This guide will get you up and running with KladML in under 5 minutes.

Installation

# Core library (lightweight, no UI)
pip install kladml

# Full CLI with Terminal UI
pip install "kladml[cli]"

Verify Installation

kladml version
# KladML version X.X.X

Option 1: Universal Quickstart

Note: The quickstart command is currently under maintenance. Please use the interactive workflow below.

Launch the Terminal User Interface for a guided experience:

kladml ui

Option 3: Traditional Workflow

Initialize a Project

kladml init

This creates the standard directory structure:

data/
├── kladml.sqlite        # Local database
├── configs/             # YAML configurations
├── datasets/            # Your data
└── projects/            # Training results
    └── {project}/
        └── {run_id}/
            ├── config.yaml
            ├── checkpoints/
            ├── exports/
            └── evaluations/

Train with Config

kladml train --config data/configs/my_experiment.yaml

Example config:

project: my-project
experiment: baseline_v1

dataset: my_data/processed
architecture: TransformerAutoencoder
params:
  d_model: 64
  n_heads: 4

training:
  epochs: 50
  batch_size: 128

export:
  auto: true
  format: onnx

evaluation:
  auto: true
  evaluator: AnomalyEvaluator

Evaluate a Run

kladml eval --run run_001 --evaluator AnomalyEvaluator --plots cdf,loglog

Compare Runs

kladml compare --runs run_001,run_002 --metric val_loss

Hyperparameter Tuning

Use Optuna integration for automated hyperparameter search:

kladml tune --config config.yaml --n-trials 50 --timeout 3600

Create Custom Models

from kladml import BaseModel, MLTask

class MyModel(BaseModel):

    @property
    def ml_task(self):
        return MLTask.CLASSIFICATION

    def train(self, X_train, y_train=None, **kwargs):
        # Your training logic
        return {"accuracy": 0.95}

    def predict(self, X, **kwargs):
        return predictions

    def evaluate(self, X_test, y_test=None, **kwargs):
        return {"accuracy": 0.93, "f1": 0.91}

    def save(self, path: str):
        # Save model artifacts
        pass

    def load(self, path: str):
        # Load model artifacts
        pass

Register it:

kladml registry register --name MyModel --path my_model.py --type model

Then use it:

kladml train --config config.yaml  # config references "MyModel"

Next Steps