Spleeter won't produce the output in .mp3 in Windows nor in Linux after the latest update

- [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

Spleeter won't produce the output in `.mp3` in Windows nor in Linux. Still, It works for `.flac`and `.wav` file formats.

## Step to reproduce

1. Installed dependencies of `ffmpeg` using `conda` and run `conda update --all` and Installed Spleeter using updated `pip` and python 3.7/8/9 (tested on different versions)
2. Using 30 GB of virtual memory, I run in `conda prompt` as ` spleeter separate --verbose -p spleeter:2stems-16kHz -f {filename}-{instrument}.{codec} -d "7200" -c "mp3" -b "320k" --mwf -B "tensorflow" -o "output_mwf" "C:\Users\MohammedMehdiTBER\Music\example.mp3" `
3. Got nothing after minutes of processing.

## Output

```DOS
(base) C:\Users\MohammedMehdiTBER\Music\Spleeter> spleeter separate --verbose -p spleeter:2stems-16kHz -f {filename}-{instrument}.{codec} -d "7200" -c "mp3" -b "320k" --mwf -B "tensorflow" -o "output_mwf" "C:\Users\MohammedMehdiTBER\Music\example.mp3"

(base) C:\Users\MohammedMehdiTBER\Music\Spleeter>
```
```ZSH
INFO:tensorflow:Apply unet for vocals_spectrogram
WARNING:tensorflow:From /home/mohammedmehditber/venv/lib/python3.8/site-packages/tensorflow/python/keras/layers/normalization.py:534: _colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Colocations handled automatically by placer.
INFO:tensorflow:Apply unet for drums_spectrogram
INFO:tensorflow:Apply unet for bass_spectrogram
INFO:tensorflow:Apply unet for other_spectrogram
INFO:tensorflow:Restoring parameters from pretrained_models/4stems/model
zsh: killed spleeter separate --verbose -B "librosa" -p -o "output_mwf" -f -d 7200 -c
```

## Environment

| OS | Window / Linux |
| Installation type | Conda / pip |
| RAM available | 6G Physical 30GB Virtual |
| Hardware spec | Intel Core i5 gen 2 |

Start at the `spleeter separate` command shown in the reproduction steps and rerun it while checking the verbose output and whether the process is killed. Compare the `.mp3` path with the working `.flac` and `.wav` cases on the reported Windows and Linux environments; done means the command produces the expected MP3 stem files.

Written by the indexing model from the issue text.