tonyhu

@FireballDWF I am taking a look at the test failure issue and trying to reproduce it locally

I've updated the Docker image to run natively on Torch 2.7.1. Meanwhile, I've fixed the test cases to account for sensitive numeric differences on short training samples

Thanks for reporting the issue! I couldn't reproduce it locally on Ubuntu with A100 GPUs using the below environment configuration and pip install. Could you try again with a different...

Hi @ncclementi, by default, it's set to 4.6 for me, but I changed it to `lightgbm==4.5` to align with the environment configuration you provided. Let me know if you need...

@ncclementi see steps below with pip freeze results of all packages: ``` python3.12 -m venv test source test/bin/activate pip install autogluon==1.3.1 --no-cache-dir ``` Feel free to dump following `pip freeze`...

I've pushed version 1.3.1 of the Docker images to the hub [here](https://hub.docker.com/repository/docker/autogluon/autogluon). Please pull the image locally to test the issue in an isolated docker environment. Let me know if...

@ncclementi Let us know if there is any further investigation required. Close this for now and feel free to re-open if help is needed

Thanks for the detailed analysis and all the useful context provided. Shout out to @jameslamb @ncclementi and @betatim. We will take a look and see if we can squeeze in...

Thank you for flagging this issue. Over the past few days, we have received several reports of failures during the 'pip install autogluon' process across various platforms. We are currently...

We have cut an issue to pip team to see if they can provide some guidance here on whaat has been changed recently: https://github.com/pypa/pip/issues/12990