diff options
author | 2017-09-06 12:03:25 -0700 | |
---|---|---|
committer | 2017-09-06 12:11:09 -0700 | |
commit | 250b3a1f543045a5de9ee7df563db09461a13e21 (patch) | |
tree | 6bc56d34facab356daa0cbb2e872f60a9ece63e1 | |
parent | 8d309ba540deea91a826ff5d81c521beec325a52 (diff) |
Improved documentation about eval ops in EstimatorSpec.
PiperOrigin-RevId: 167753099
-rw-r--r-- | tensorflow/python/estimator/model_fn.py | 5 |
1 files changed, 4 insertions, 1 deletions
diff --git a/tensorflow/python/estimator/model_fn.py b/tensorflow/python/estimator/model_fn.py index 1a4b0c5fc0..cfa4be5c7d 100644 --- a/tensorflow/python/estimator/model_fn.py +++ b/tensorflow/python/estimator/model_fn.py @@ -131,7 +131,10 @@ class EstimatorSpec( train_op: Op for the training step. eval_metric_ops: Dict of metric results keyed by name. The values of the dict are the results of calling a metric function, namely a - `(metric_tensor, update_op)` tuple. + `(metric_tensor, update_op)` tuple. `metric_tensor` should be evaluated + without any impact on state (typically is a pure computation results + based on variables.). For example, it should not trigger the `update_op` + or requires any input fetching. export_outputs: Describes the output signatures to be exported to `SavedModel` and used during serving. A dict `{name: output}` where: |