How to modify SOM_LM-SNNs.py for batch training?

[AXYZdong]

I noticed that batch_size = 1 in SOM_LM-SNNs.py and if I increase the batch_size it will report an error.

[Hananel-Hazan]

Although the code can be adapted to different batch_size, I dont know if the algorithm learning method can support higher batch_size. You are welcome to adapt the code to support higher batch_size, let me know how higher batch_size effect the learning, this is interesting! I suspect that this kind of change need to change the inhibition mechanism too.

[AXYZdong]

I want to increase batch_size to fully utilize the GPU. I try to adapt the code to support higher batch_size, but I've had no success so far.

[Hananel-Hazan]

The update rule and the inhibition signal need to be updated too.