Exporting band composites in the right intensity

I am trying to export different band composites such as the True Color (RGB), Color Infrared (NIR R G), Short-Wave Infrared (B12, B8A, B4), and Agriculture (B11, B8, B2) to JPEG image format.
Exporting the True Color patches was using the following code snippets where I specified (as shown in eo-learn examples): bands=np.clip(eopatch.data['BANDS'][closest_date_id][..., [3, 2, 1]] * 3.5, 0, 1).the intensity levels (I am guessing) to 3.5, 0, 1.
My question is how can I do the same for the other band composites what intensity levels should I use? or should I use the same values? i.e: 3.5, 0, 1.

Here you are multiplying your bands by 3.5 for visualisation, and clipping the range to the 0-1 interval. Numpy clip documentation.

It all depends on what you want to do with the output. I suggest you experiment by changing the 3.5 value and see how it affects your results.