Running trained models on different data

[Qwlouse]

I've reproduced the results and verified that they are fine. However I'd like to run the trained models on some new data, and I can't seem to make that work. Maybe you could point me in the right direction? Is it at all possible without retraining everything, because to me it seems you didn't store any trained models, only their predictions.

[alexandrebarachant]

Hi Klaus,

The bad news is that we did not save the models so you will have to re-run things.
The good news is that you can keep validation predictions and therefore have to re-run only the test predictions.

Before getting to into the how to generate and save models, i strongly suggest to choose the "safe1" solution. It only involves a subset of 8 level2 models, and the performance does not change that much. It will be much faster to run.

Most of the change are concerning level1 models. I assume your new data are in the same format as the other ones, so here we go :

• Update meta-info, in genInfos.py, you will have to point out to your new files (L33).
• Re-run genInfos.py
• In preprocessing/aux.py , change the path L44 to point the new files.
• do the same thing for genPreds_CNN_Tim.py and genPreds_RNN.py
• re-run lvl1 models in test mode only (modify lvl1/genAll.sh)
• re-run lvl2 models in test mode only (modify lvl2/genSafe1.sh)

If you want to save models, you could use the guideline provided by sklearn : http://scikit-learn.org/stable/modules/model_persistence.html

For lvl1, you will have to save pre-processing model, post-preprocessing, and final classifyer. for lvl2, you just need to save the classifyer

[ut33]

When I try to run the code, I get an IndexError: list index is out of range
Is anyone else getting this problem? What should I do?

---

Traceback (most recent call last):
File "genInfos.py", line 35, in
raw_val = concatenate_raws([creat_mne_raw_object(fname, read_events=True) for fname in fnames_val])
File "", line 2, in concatenate_raws
File "C:\ProgramData\Anaconda3\envs\mne\lib\site-packages\mne\utils.py", line 729, in verbose
return function(*args, **kwargs)
File "C:\ProgramData\Anaconda3\envs\mne\lib\site-packages\mne\io\base.py", line 2500, in concatenate_raws
raws[0].append(raws[1:], preload)
IndexError: list index out of range

[kritigoyal]

@ut33 were you able to figure out how to resolve this issue