Welcome to the Linux Foundation Forum!

Linux, Python, C ++, Brain2 and errors.

Options
piotrbujakowski
piotrbujakowski Posts: 1
edited June 25 in Software Development

Hi.

I try to make and run first time AI Neural Network on my computer.
Knowledge Python is basic.

I'm using Linux Ubuntu.

I install whole in system using apt.
When show communicate about pip, then I'm doing that.
[Is a lot of problems with right file names and compatibility.]

I try to configure to direction in home catalogue [advice from Gemini].
Port 65432 is unblocked in firewall.

Helpful is Gemini, but can't solve problems what I have now.

Please help me with that.

Python engine:

import brian2
import socket
import numpy as np

from brian2 import Neuron, Eq

Define neuron equations

eqs = '''
dv/dt = (I - gl(v-Vl) - gNa(m3)h(v-VNa) - gK*(n4)(v-VK)) / Cm
dm/dt = alpha_m(v)
(1-m) - beta_m(v)m
dh/dt = alpha_h(v)
(1-h) - beta_h(v)h
dn/dt = alpha_n(v)
(1-n) - beta_n(v)*n
'''

Define parameters and functions for the equations (replace with actual definitions)

Vl = -65mV # Leak reversal potential
VNa = 50
mV # Sodium reversal potential
VK = -70mV # Potassium reversal potential
Cm = 1
uF/cm2 # Membrane capacitance
gl = 0.3e-3siemens/cm2 # Leak conductance
gNa = 120e-3
siemens/cm2 # Sodium conductance
gK = 36e-3*siemens/cm2 # Potassium conductance

Define neuron object using custom equations

neurons = Neuron(eqs=eqs, methods={'alpha_m': alpha_m, 'beta_m': beta_m, 'alpha_h': alpha_h, 'beta_h': beta_h, 'alpha_n': alpha_n, 'beta_n': beta_n}) # Include all required methods

Standard variables

num_neurons = 1000
duration = 1000 # Simulation duration in milliseconds

(Replace with your implementation)

Hodgkin-Huxley neuron model with STDP learning (replace with your specific neuron and synapse definitions)

... (Include your specific code for defining synapses and learning rules)

Data variables

average_firing_rates = []
average_synaptic_weights = []

Recording functions

def record_firing_rates():
global average_firing_rates
average_firing_rates.append(np.mean(spike_monitor.count / (duration * 1000))) # Convert to Hz

def record_synaptic_weights():
global average_synaptic_weights
weights = synapses.weight # Assuming you have a 'synapses' object with weight attribute
average_synaptic_weights.append(np.mean(weights))

Network monitors

spike_monitor = brian2.SpikeMonitor(source=neurons)
brian2.NetworkOperation(record_firing_rates, dt=10brian2.ms) # Record every 10 ms
brian2.NetworkOperation(record_synaptic_weights, dt=100
brian2.ms) # Record every 100 ms

Network socket setup (replace with your IP address and port)

HOST = '127.0.0.1' # Standard loopback interface address (localhost)
PORT = 65432 # Port to listen on (non-privileged ports are > 1023)
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.bind((HOST, PORT))

Run the simulation

brian2.run(duration * brian2.ms)

Send data to C++ program (replace with IPC if on the same machine)

data = {'firing_rates': average_firing_rates, 'synaptic_weights': average_synaptic_weights}
data_str = str(data) # Convert data to string for sending
s.sendall(data_str.encode())

Close the socket

s.close()

I have to second code in C++:

include

include <stdio.h>

include <stdlib.h>

include <unistd.h>

include <string.h>

include <sys/types.h>

include <sys/socket.h>

include <netinet/in.h>

include <arpa/inet.h>

include

include <jsoncpp/json.h> // Include JSON library for parsing

int main() {
// Standard variables
int port = 65432; // Port used by Brian 2 simulation
std::string host = "localhost"; // Replace with IP address of Brian 2 (if not localhost)

// Socket setup
int sockfd;
struct sockaddr_in servaddr;

sockfd = socket(AF_INET, SOCK_STREAM, 0);
if (sockfd == -1) {
perror("socket creation failed");
exit(EXIT_FAILURE);
}

memset(&servaddr, 0, sizeof(servaddr));

servaddr.sin_family = AF_INET;
servaddr.sin_port = htons(port);
servaddr.sin_addr.s_addr = inet_addr(host.c_str());

if (connect(sockfd, (struct sockaddr*)&servaddr, sizeof(servaddr)) != 0) {
perror("connection failed");
exit(EXIT_FAILURE);
}

// Data receiving and processing loop
while (1) {
char buffer[1024]; // Adjust buffer size based on data volume
int n = recv(sockfd, buffer, sizeof(buffer), 0);
if (n == 0) {
printf("Connection closed by server\n");
break;
} else if (n == -1) {
perror("recv failed");
exit(EXIT_FAILURE);
}

// Parse received data (assuming JSON format)
std::string data_str(buffer, n);
Json::Reader reader;
Json::Value data;
if (!reader.parse(data_str, data)) {

Error:

Python 3.12.3 (main, Apr 10 2024, 05:33:47) [GCC 13.2.0] on linux
Type "help", "copyright", "credits" or "license()" for more information.

= RESTART: /home/peter/python/ai_neural_brain_2/engine_brain_2.py =
ERROR Brian 2 encountered an unexpected error. If you think this is a bug in Brian 2, please report this issue either to the discourse forum at http://brian.discourse.group/, or to the issue tracker at https://github.com/brian-team/brian2/issues. Please include this file with debug information in your report: /tmp/brian_debug_vkzoele7.log Additionally, you can also include a copy of the script that was run, available at: /tmp/brian_script_fgiuug_k.py Thanks! [brian2]
Traceback (most recent call last):
File "/usr/lib/python3.12/idlelib/run.py", line 580, in runcode
exec(code, self.locals)
File "/home/peter/python/ai_neural_brain_2/engine_brain_2.py", line 5, in
from brian2 import Neuron, Eq
ImportError: cannot import name 'Neuron' from 'brian2' (/usr/lib/python3/dist-packages/brian2/init.py)

Categories

Upcoming Training