Add Stroop Model to synthetic models

[TheLemonPig]

Added stroop_model.py to synthetic models

[TheLemonPig]

Where do we add package dependencies? In this case, the error is caused by not recognizing torch

[younesStrittmatter]

package requirements are added to the dependency section in the pyproject.toml file

[hollandjg]

Is there a strong reason to use PyTorch for this? If we're just doing matrix transformations, numpy should be fine, and is already part of the dependencies. Torch is great for complicated neural network things, but if it's possible to simplify this so it works with numpy alone, that would be better. (PyTorch is a big dependency which we've had trouble with in the past, so it's best to avoid it if possible.)

[TheLemonPig]

Update to match Younes' PR changes

[benwandrew]

this looks good, including updates to match Younes' changes, but i don't know about @hollandjg's comment re: PyTorch... ?

[TheLemonPig]

Is there a strong reason to use PyTorch for this? If we're just doing matrix transformations, numpy should be fine, and is already part of the dependencies. Torch is great for complicated neural network things, but if it's possible to simplify this so it works with numpy alone, that would be better. (PyTorch is a big dependency which we've had trouble with in the past, so it's best to avoid it if possible.)

This is a good point. This would mean changing work done by @musslick 's code, which feels outside of my domain without direction.

[TheLemonPig]

Attempting to replace torch with numpy, with help of chatGPT

[benwandrew]

Attempting to replace torch with numpy, with help of chatGPT

how'd it go!? superficially looks good, but curious about any testing you performed (or have planned).

[TheLemonPig]

I haven't done any testing yet. I was under the assumption the prebuilt tests checked the functionality of all models. I am actually quite surprised that all the test passed on the first attempt XD. The only additional thing I can think to test is whether the new model replicates the old results. Unfortunately, I have a poor understanding of how this and the previous model worked and what its results were. I am not sure how I would proceed with that

[younesStrittmatter]

Please wait for #10 and then change accordingly

[benwandrew]

Please wait for #10 and then change accordingly

@TheLemonPig i think we will also need to reflect this change in the synthetic models you already merged (yesterday).

[musslick]

Looks great! Might just want to check if it still runs with Younes' suggested changes.

[benwandrew]

looks good, as long as it runs with recent changes!