Spleeter simply doesn't work

- [x] I didn't find a similar issue already open.
- [x] I read the documentation (README AND Wiki)
- [x] I have installed FFMpeg
- [x] My problem is related to Spleeter only, not a derivative product (such as Webapplication, or GUI provided by others)

## Description

After installing Spleeter with pip, it tried to use it. At the beginning, it gave some errors, saying that it wasn't able to load the path because "it was None". I searched on the "Issue" tab of this project and found out that, apparently, it was caused by the ```pretrained_models/<2-4-5>stems``` folders being empty, and Spleeter wasn't able to "fill" them. I manually downloaded and extracted the content of those folders, and if I try to execute the commands, two possible things can happen:

1. The program doesn't print anything on the terminal, giving the user the possibility to type another command. (happened with 5stems and 4stems)
2. The program still doesn't print anything, remaining completely frozen in its state. While it consumes a lot of RAM and CPU in the beginning, it slowly goes down, still remaining frozen. (happened with 2stems)

## Step to reproduce

I simply typed ```python -m spleeter separate -o audio_output -p spleeter:5stems audio_example.mp3```.
Then I tried changing ```5stems``` to ```4stems``` and ```2stems```. Same result as before.

## Environment

| | |
| ----------------- | ------------------------------- |
| OS | Windows 11 |
| Installation type | pip |
| RAM available | 8GB |
| Hardware spec | Intel(R) Core(TM) i5-1035G1, iUGPU |

## Additional context

Start by reproducing the reported `python -m spleeter separate` command on the stated Windows 11, pip-installed environment with each of the 2stems, 4stems, and 5stems models. Trace the `spleeter separate` entry point and model-loading path; done means the failure or hang has a confirmed cause and a reproducible fix.

Written by the indexing model from the issue text.