How to use the python bridge for a machine learning strategy

From the manual: "You can use [Train] mode and WFO for training Python machine learning algorithms, f.i. a SVM or neural network. For this, send training data to Python at the end of every WFO cycle, and store the trained models in files for later use in the [Test] or [Trade] session, similar to the neural function for R. A neural function for Python will be provided in a future Zorro version. "

I am trying to understand the best way to send the training data to Python at the end of the training period.

Would I append the signals for each bar to some file on every loop of run()? Or is there some way obtain the entire signals data set over the entire training run to be passed on to Python?

You normally train the entire data set at the end of any wfo cycle. Check out the neural function in r.h - you'll need the same function, only calling Python code instead of R code.

So I write my own function neural, if I understand. Makes sense.

I thought using python for deep learning isn't possible yet, because all python machine learning packages are 64bit only and there is no python64bit bridge yet..???

I never heard that, and I have installed tensorflow on my 32-bit python installation.

It looks like you were correct.

I was using python from the R bridge. I assumed it was 32-bit, but I must have been mistaken. I wish someone had corrected me.

So we can only use 64-bit python with Zorro through the R-bridge?? That seems ridiculous. I will have to check the forum for discussion on the python bridge.

Here is the relevant thread: [url=https://opserver.de/ubb7/ubbthreads.php?ubb=showflat&Number=477133]#477134