Upload a file, or let the Python client convert your model. Either way, MLGatee installs the versions your model needs.
Models and pipelines, including pandas column transformers. MLGatee installs the scikit-learn version your file was saved with.
mlgatee.deploy(model, name="iris")Pickled models get the exact version that saved them, and XGBoost's own .json and .ubj files work too. Installed without about 345 MB of GPU code.
mlgatee.deploy(booster, name="churn")Classifiers and regressors saved with pickle or joblib.
mlgatee.deploy(model, name="scores")Classifiers and regressors, including table models that mix text and number columns. CatBoost builds take a little longer.
mlgatee.deploy(model, name="pricing")Read without loading and checked for custom operators and external data. Classifiers also answer /predict_proba.
mlgatee.deploy("model.onnx", name="iris")Float, float16 and int8 models. A file that needs full TensorFlow is refused with the name of the op it needs.
mlgatee.deploy("model.tflite", name="digits")The Python client converts your Keras model to TensorFlow Lite in your own Python, then checks it gives the same answers before uploading.
mlgatee.deploy(keras_model, name="digits")The Python client exports ONNX, runs both versions on your example input and compares the answers before uploading.
mlgatee.deploy(net, name="digits", example_input=X[:5], task="classifier")Small text models for classification or embeddings, packed with their tokenizer into one 8-bit file. Send plain text to /predict.
mlgatee.deploy_hf("philschmid/tiny-bert-sst2-distilled", name="reviews")Native TensorFlow and Keras, native PyTorch and full Hugging Face Transformers need more than 512 MB just to start. They’re planned for the Large machine add-on. Until then, convert them with the Python client. Image models take numbers: do the preprocessing in your own code.
Select a framework to see how it deploys.
Models and pipelines, including pandas column transformers. MLGatee installs the scikit-learn version your file was saved with.
mlgatee.deploy(model, name="iris")installs: scikit-learn at the version saved in your fileSelect a framework to see how it deploys
Upload a file or deploy from your notebook. MLGatee handles the packages, the server and the key.
Free plan: one model, no card needed