experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Print output of predictions in simple estimator example for DistributionStrategy

Browse Source 
Resolved the TODO that asked to print meaningful results from the Estimator's predictions.
Collects all elements yielded by the generator in a list and prints that list instead of the generator object.
Results in the following output: `Prediction results: [{'logits': array([1.0162734], dtype=float32)}, ..., {'logits': array([1.0162734], dtype=float32)}]`.
This commit is contained in:
Joppe Geluykens 2018-08-18 13:46:56 +02:00 committed by GitHub
parent a6d62ffea3
commit 93e2a37a5f
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
tensorflow/contrib/distribute/python/examples/simple_estimator_example.py
View File

@ -78,9 +78,8 @@ def main(_):
predict_features = tf.data.Dataset.from_tensors([[1.]]).repeat(10)
return predict_features
predictions = estimator.predict(input_fn=predict_input_fn)
# TODO(anjalsridhar): This returns a generator object, figure out how to get
# meaningful results here.
prediction_iterable = estimator.predict(input_fn=predict_input_fn)
predictions = [prediction_iterable.next() for _ in range(10)]
print("Prediction results: {}".format(predictions))