TensorFlow upgrades break working setups

2 posts from 2 people2d activeSteady, +1 this weekunverified

Posts per day2 posts, Nov 6 to Nov 13

The posts behind it

GitHub issuesmlflow/mlflow · mlflow-app[bot]Nov 8

Flaky `IntegrityError` on `logged_model_metrics` in `tests/keras/test_autolog.py::test_default_autolog_behavior` (async metrics vs. `log_model` race)

> [!NOTE] > Created by the [Nightly workflow run](https://github.com/mlflow/mlflow/actions/runs/42491372153/attempts/1). ### Summary `tests/keras/test_autolog.py::test_default_autolog_behavior` failed intermittently with `sqlite3.IntegrityError: UNIQUE constraint failed:...

2 replies

GitHub issuesmlflow/mlflow · fruitstashNov 6

[BUG] mlflow.keras.autolog modifies the user's callbacks list, so a second fit() logs a FAILED run plus an extra RUNNING run

### PR policy - [x] I understand that PRs for this issue are closed automatically unless a maintainer has applied the `ready` label, the issue isn't assigned to someone else, and no open PR already links it. - [x] I understand a triage bot may open a PR for this issue...

2 replies

Companies and products named

TensorFlow

2 posts

MLflow

2 posts

SQLite

2 posts

GitHub

2 posts

About this problem

Evidence

2 posts from 2 people in 1 place, about 2 a week over 3 days. Mostly on mlflow/mlflow. Tools named alongside: MLflow, SQLite, GitHub.

Frustration Frustration 0 of 3· Seen on GitHub issues

How it was grouped

Posts that state a pain and match the "breaking-changes" rule about TensorFlow. First post Nov 6, 2026, latest Nov 8, 2026. Unverified: fewer than 3 posts, or one person or place only. Method

The same problem elsewhere