HistGradientBoosting

You can run and interact with this HistGradientBoostingClassifier MCP server to train, predict, and manage HistGradientBoostingClassifier models. It exposes handy tools for creating classifiers, training on data, generating predictions and probabilities, evaluating performance, inspecting models, and serializing or loading models for reuse.

How to use

You will connect to the MCP server using a compatible MCP client. Once connected, you can perform end-to-end machine learning workflows including creating a classifier with custom parameters, training it on your data, making predictions and probability estimates on new samples, evaluating accuracy, and inspecting or managing stored models. The server stores models in memory for the current run, so expect persistence only for the active session unless you add persistent storage.

How to install

Prerequisites: you need Python and access to install project dependencies. Ensure you have a working Python environment and permissions to install packages.

pip install -r requirements.txt

Additional notes

Configuration and deployment details are described here to help you run and connect to the MCP server in your environment. You will find available tools to create, train, predict, and manage HistGradientBoostingClassifier models, along with guidance on model storage and how to verify your deployment.

Configuration and deployment

The server provides a public HTTP endpoint for remote access. You can connect using an MCP client and the provided URL, then invoke tools such as creating classifiers, training models, predicting, and evaluating performance.

Security and persistence notes

Currently, models are stored in memory during the server lifetime. Restarting the server will lose all models. For production use, consider implementing persistent storage such as a database, file system, or cloud storage.

Available tools

create_classifier

Create a new HistGradientBoostingClassifier with custom parameters such as learning_rate, max_iter, and max_leaf_nodes.

train_model

Train a classifier on provided feature data X and label vector y.

predict

Produce class predictions for new data samples X.

predict_proba

Return class probability estimates for input samples X.

score_model

Evaluate model accuracy or other metrics on a test dataset.

get_model_info

Retrieve detailed information about a stored model, including parameters and training metadata.

list_models

List all available models currently stored in memory.

delete_model

Remove a model from memory by its model_id.

save_model

Serialize a model to a base64 string for storage or transfer.

load_model

Load a model from a serialized base64 string.