Experiment tracking for Machine and Deep learning

ModelChimp is an experiment tracking tool that helps machine learning engineers log changes, compare experiments, track hyperparameter changes, and analyze results in real time.

Frameworks supported
With a host of features built into ModelChimp, we assist developers with the right tools that eliminates the grunt work associated with model building


Track experiments, metrics, and hyperparameter changes on a real-time basis. Whether you're running experiments using a script or a Jupyter Notebook, we've got you covered.

           Tracker(key='<key>', host='app.modelchimp.com',
experiment_name="MINST B")


Records metrics across experiments and epochs. Visualize the metric history for each experiment or epoch via a chart. Use Grid Search to find the right set of hyperparameters for a particular model.

            tracker.add_param("Accuracy_Train" = train_accuracy)


Store associated data from images to the models themselves with our Data Store. Record meta information associated with every file. Analyze the impact of data on every experiment.

tracker.add_asset(<filename>, {'accuracy': accuracy})
           tracker.add_asset(<filename>, {'accuracy': accuracy})


Share code, metrics, hyperparameters changes with fellow team members. Collaborate over the model building process with a click of a button.

Experiment tracking for the community

ModelChimp is a 100% opensource project. Clone our repo from GitHub & start tracking experiments within minutes.

Schedule a Demo