These functions are used to preprocess and postprocess inputs and outputs of Keras applications.
Usage
application_preprocess_inputs(model, x, ..., data_format = NULL)
application_decode_predictions(model, preds, top = 5L, ...)Arguments
- model
A Keras model initialized using any
application_function.- x
A batch of inputs to the model.
- ...
Additional arguments passed to the preprocessing or decoding function.
- data_format
Optional data format of the image tensor/array.
NULLmeans the global settingconfig_image_data_format()is used (unless you changed it, it uses"channels_last"). Defaults toNULL.- preds
A batch of outputs from the model.
- top
The number of top predictions to return.
Value
A list of decoded predictions in case of
application_decode_predictions().A batch of preprocessed inputs in case of
application_preprocess_inputs().
Functions
application_preprocess_inputs(): Pre-process inputs to be used in the modelapplication_decode_predictions(): Decode predictions from the model
Examples
if (FALSE) {
model <- application_convnext_tiny()
inputs <- random_normal(c(32, 224, 224, 3))
processed_inputs <- application_preprocess_inputs(model, inputs)
preds <- random_normal(c(32, 1000))
decoded_preds <- application_decode_predictions(model, preds)
}