Define an Estimator Specification
Define the estimator specification, used as part of the model_fn
defined with
custom estimators created by estimator()
. See estimator()
for more details.
estimator_spec(mode, predictions = NULL, loss = NULL, train_op = NULL, eval_metric_ops = NULL, training_hooks = NULL, evaluation_hooks = NULL, prediction_hooks = NULL, training_chief_hooks = NULL, ...)
mode |
A key that specifies whether we are performing
training ( |
predictions |
The prediction tensor(s). |
loss |
The training loss tensor. Must be either scalar, or with shape |
train_op |
The training operation – typically, a call to |
eval_metric_ops |
A list of metrics to be computed as part of evaluation.
This should be a named list, mapping metric names (e.g. |
training_hooks |
(Available since TensorFlow v1.4) A list of session run hooks to run on all workers during training. |
evaluation_hooks |
(Available since TensorFlow v1.4) A list of session run hooks to run during evaluation. |
prediction_hooks |
(Available since TensorFlow v1.7) A list of session run hooks to run during prediciton. |
training_chief_hooks |
(Available since TensorFlow v1.4) A list of session run hooks to run on chief worker during training. |
... |
Other optional (named) arguments, to be passed to the |
Other custom estimator methods: estimator
,
evaluate.tf_estimator
,
export_savedmodel.tf_estimator
,
predict.tf_estimator
,
train.tf_estimator
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.