mlflow / mlflow

Open source platform for the machine learning lifecycle

Home Page:https://mlflow.org

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

[BUG]

esemichasnova opened this issue · comments

Willingness to contribute

No. I cannot contribute a bug fix at this time.

MLflow version

2.11.3

System information

  • OS Platform and Distribution (e.g., Linux Ubuntu 16.04):
  • Python version:
  • yarn version, if running the dev UI:

Describe the problem

When I set a alias to the existed model and the alias contain capital letters, it converts into the alias without capital letters.
E.g. I am typing "CapitalLetterTag" and the appeared alias is "capitallettertag"

image image

Steps to reproduce the bug

ezgif-7-f96cf66172

Code to generate data required to reproduce the bug

No response

Is the console panel in DevTools showing errors relevant to the bug?

No response

Does the network panel in DevTools contain failed requests relevant to the bug?

No response

@esemichasnova Thank you for opening the issue. This is by-design, model alias feature is not case-sensitive. For readability, you can use hyphens instead of capital cases e.g. capital-letter-tag. We will update the documentation to make this behavior clear!

Hi, just to clarify the scenario here:

  1. When we create the alias MyNewAlias via an API call, then it works fine and we see it as MyNewAlias in the MLflow frontend. When we use the API to search and list model versions' aliases, we also see MyNewAlias.

  2. When we create an alias from the frontend (clicking the pencil next to the aliases box) and input MyNewAlias, we see that it is saved as mynewalias.

This seems like a bug to me.

@BalkanFlink Thank you for the additional information!

Sorry for the misinformation here. After further investigation, it turns out the alias is not case sensitive only in particular environments (e.g. Databricks, local Mac OS with file store). For other environments, the alias might be case sensitive.

We will work on fixing the UI and update you soon.

@mlflow/mlflow-team Please assign a maintainer and start triaging this issue.