Skip to content

Releases: sayakpaul/ml-deployment-k8s-fastapi

ONNX optimized ResNet50 and ResNet50 from `keras.applications`

21 Feb 02:19
2721029
Compare
Choose a tag to compare
add notebook for tf to onnx conversion (#1)

* add notebook for tf to onnx conversion

* update: use default ResNet50, compare tf.keras vs ONNX converted model on CPU

* update: use built-in preprocess_input(), model.predict method