Matti Picus

From the traceback, it seems like ray is trying to recover from an error around the use of `edge_detect_time`, which is, as far as I can tell, not a ray...

@alexsaurber would it be possible to get a reproducer and the traceback?

ray may need a [hook for pyinstaller](https://pyinstaller.org/en/stable/hooks.html). Could you provide a more complete reproducer in the form of commands to create the conda environment, a demo application, and the command...

What version of CUDA are you using?

I could not reproduce when using the resource settings below, the second script ran to completion. I have a windows 10 machine with 64GB RAM. `resources_per_trial={"cpu": 6, "gpu": 0, "memory":...

I created the environment via ``` >conda create -n issue27646 python=3.9 >conda activate issue27646 >conda install --channel pytorch pytorch=1.11 pip pandas >pip install "ray==1.13" "ray[tune]" ```

I tried to run the first script. After I run `conda install -c pytorch torchvision=0.12` I needed to set `resume="AUTO"`, and then after a while I get `OSError: [WinError 1455]...

@richardliaw any thoughts?

Using ray==2.0.0rc0 still crashes after ~124 trials. The information is saved, so restarting continues where it left off.

Yes, it exits unexpectedly for me too after it saves the data and prints out a summary.