PG.

PG.

PG.

PG.

Allow to have a good idea of the program functionning.
Display stays on the same line using \r and a flush of stdout

PG.

In blind mode there is no keyboard control so step by step mode is meaningless.

PG.

Data saved ibn promethe from : f_save_image_activity() or f_display_image_activity()

PG.

PG.

When the input is higher than a given threshold the f_cond_breakpoint function switch to the step by step mode.
After a continue, the f_cond_breakpoint is not active anymore until the input goes below the threshold.
The function only reacts to the triggering on of the input (rising front).

Do not forget to add the new source in the list of the files to be compiled ;-).

PG.

PG.

This can be useful to follow the execution of a long script.

PG.

PG.

PG.

The old one should be suppressed in a next commit if everything is working well.

Add an option -b1 to save in binary files to improve performances. Default is still ASCII format for compatibility.
PG.

Update applications/example/timing_learning/README.md

See merge request !1

PG.

PG.

Now all weights of all the neurons nearby the click point are displayed.
It solves the issue of clicking on the right micro neurons when the group as a lot of neurons.

PG.

The modulo property is now functinig correctly.
Suppress also some debug information.

PG.

PG.

PG.

PG.

PH.

Add comment and modif printf. This file need to be modified after coeff->val will not be used anymore.
PG.

Midify source code so that the old function with coeff->val are still saved correctly until their change.
use to copy_back function to put coeff-> val values in coeffs_val[] array.

PG.

PG.