Tests keep failing even though HF space and colab notebook are fine

[sgbaird]

https://github.com/sparks-baird/self-driving-lab-demo/actions/runs/11611396522

And

https://github.com/sparks-baird/self-driving-lab-demo/actions/runs/11611145405/job/32331919350#step:5:207

show that there's some stochasticity to the tests (and that testing the public demo can complete successfully).

[sgbaird]

@SissiFeng any thoughts here? Would you mind helping with troubleshooting this a bit? Debugging with GitHub Actions is a bit slower, but I think I was able to run all the tests locally without issue. So, I think it's something more specific to the way things are running on gh actions.