Improving Sample Efficiency of Value Based Models Using Attention and Vision Transformers