onnxmltools and tensorflow-onnx
These are complementary tools that together enable ONNX conversion from different ML frameworks—ONNXMLTools handles scikit-learn, XGBoost, and LightGBM models while TensorFlow-ONNX specifically converts TensorFlow/Keras/TFLite models, so users typically choose the appropriate converter based on their source framework rather than selecting between them.
About onnxmltools
onnx/onnxmltools
ONNXMLTools enables conversion of models to ONNX
This tool helps machine learning engineers and data scientists convert their trained models from various frameworks like scikit-learn, TensorFlow, or Core ML into the ONNX format. You provide a model trained in one of the supported toolkits, and it outputs an equivalent model in the ONNX standard format. This allows for easier deployment and portability across different inference runtimes and hardware.
About tensorflow-onnx
onnx/tensorflow-onnx
Convert TensorFlow, Keras, Tensorflow.js and Tflite models to ONNX
This tool helps machine learning engineers and data scientists convert their trained models built with TensorFlow, Keras, TensorFlow.js, or TFLite into the ONNX format. You provide your existing model file, and it outputs a universal ONNX model, which can then be deployed across various hardware and runtimes. It's for anyone needing to standardize or port their TensorFlow ecosystem models for broader compatibility and deployment flexibility.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work