ORNL / flowcept

Runtime data integration system that empowers any data processing system to capture and query workflow provenance using data observability.
MIT License
1 stars 3 forks source link

Tensorboard interceptor keeps receiving events even after training #49

Open renan-souza opened 1 year ago

renan-souza commented 1 year ago

This has low priority because it can be easily worked around by calling consumer.stop(), but this is a bug that should be fixed.

renan-souza commented 1 year ago

I removed this test: https://github.com/ORNL/flowcept/blob/main/tests/plugins/test_tensorboard.py#L155 in this commit: https://github.com/ORNL/flowcept/commit/6d0ce57cd32b1148e5171e38f9c87fd40528d1e4 This test may be related to this issue. When we fix tensorboard adapter issues, we should return with this test.

renan-souza commented 7 months ago

As of today, sometimes the assert len(docs) == 16 check still causes intermittent errors. See: https://github.com/ORNL/flowcept/actions/runs/8023284655/job/21920203496?pr=111