Pass path to config file when running training_ml.py and update README

[RemiLehe]

The training_ml.py had some complicated logic to decide where to look for the config file:
https://github.com/BLAST-AI-ML/synapse/blob/main/ml/train_model.py#L70
This makes the code behavior difficult to understand/predict.

However, in practice there is only two cases that we care about:

• train_model.py run through the sbatch file training_pm.sbatch (this includes triggering the training through the dashboard, which does launch training_pm.sbatch via the SFAPI): in this case train_model.py uses the config file which is mounted into the container (/app/ml/config.yaml).

• train_model.py run locally from a terminal (python train_model.py ...), for training purposes: in this case, it is convenient to be able to directly point to the config file that we want to use for testing.

This PR accomplishes this by replacing the --experiment command line argument by --config_file_path

[EZoni]

Since the ML training README starts with titles at level # (not ##) could we maybe adjust the title levels in the dashboard README as well, to make the two consistent in formatting and style?

[EZoni]

We can do this in a separate PR in the interest of having the full testing/debugging mode feature available asap.