Regarding Inhibitory Neurons and Excitatory Neuron Trainings Under Dale's Rule #680
-
|
Hi! I'm recently trying to create a network with inhibitory neurons and excitatory neurons training under Dale's rule, where weights connecting neurons would not change their sign during training. I'm wondering if there is a way I could easily do this with some parameters passed to node or connection class. Thanks alot in advance! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
I've incorporated Dale's rule into my branch, which is a valuable addition. To utilize it, simply include the Dales_rule parameter in the connection object. This parameter accepts a Boolean matrix with the same dimensions as weights, where True represents a positive connection and False indicates a negative connection. It's important to note that implementing this rule may significantly slow down the runtime. I may work on optimizing it for better performance in the future. Feel free to try it out and let me know your thoughts. |
Beta Was this translation helpful? Give feedback.
-
|
Sure, I'll try it out recently. Thank you for your help! |
Beta Was this translation helpful? Give feedback.
I've incorporated Dale's rule into my branch, which is a valuable addition. To utilize it, simply include the Dales_rule parameter in the connection object. This parameter accepts a Boolean matrix with the same dimensions as weights, where True represents a positive connection and False indicates a negative connection. It's important to note that implementing this rule may significantly slow down the runtime. I may work on optimizing it for better performance in the future. Feel free to try it out and let me know your thoughts.