Hyperparameter tuning using talos: random search

Question

I am using talos to tune the hyperparameters of a multi input multi output neural network. I want to use randomized search; however, when use random_method option in talos, it tries each and every combination of the hyperparameters. Here are the parameters that I use :

p = {'neurons_per_layer':[225,250,275],
      'targeted_reg':[False],
      'rtio':[1.],
      'val_split':[0.25,],
      'batch_size':[1,64,128],
      'epochs':[80,100],
      'learning_rate':[0.5,0.1,0.05],
      'reg_l2':[0.5,0.1,0.05],
      'loss_func':[dragonnet_loss_binarycross]}

Here is how random search looks like in talos:

I can't see any difference between grid search and random search in talos. Where am I wrong and how can I increase the speed of tuning?