tf.estimator.experimental.call_logit_fn

TensorFlow 1 version View source on GitHub

Calls logit_fn (experimental).

tf.estimator.experimental.call_logit_fn(
    logit_fn, features, mode, params, config
)

THIS FUNCTION IS EXPERIMENTAL. Keras layers/models are the recommended APIs for logit and model composition.

A utility function that calls the provided logit_fn with the relevant subset of provided arguments. Similar to tf.estimator._call_model_fn().

logit_fn A logit_fn as defined above.
features The features dict.
mode TRAIN / EVAL / PREDICT ModeKeys.
params The hyperparameter dict.
config The configuration object.

A logit Tensor, the output of logit_fn.

ValueError if logit_fn does not return a Tensor or a dictionary mapping strings to Tensors.