tensorflow / ecosystem

Integration of TensorFlow with other open-source frameworks
Apache License 2.0
1.37k stars 392 forks source link

Distribution Strategies Example BaseCollectiveExecutor::StartAbort Out of range #138

Closed runzhliu closed 1 year ago

runzhliu commented 4 years ago

Hi, I'm testing Distribution Strategies Example on my own K8S cluster, there is an unexpected error before the end of the task.

image

However, when I run keras_model_to_estimator.py locally, it ends normally.

image