Text this: SA-MARL: Novel Self-Attention-Based Multi-Agent Reinforcement Learning With Stochastic Gradient Descent