Add all relevant (hyper)params to tensorboard

[Haydnspass]

The stuff that we put in the param.yaml should be also found in the tensorboard logger. Log it there.

[ASpeiser]

Suggestion, something like:

pdict = param.to_dict()
for k in pdict:
    if pdict[k] is not None:
        pstring = pprint.pformat(pdict[k], indent=2).replace('\n  ','  \n')
        writer.add_text(k, pstring)