sample1.py is the code I used for the mountain car problem and sample2 is for the pendulum.
The policies for sample1 and sample2 are saved in policies1 and policies2, respectively.
Each file takes about 2 minutes to train the agent. If you just want to see the output of the training you can comment the "train()" methods in each file. (Line 71 in sample1 and line 63 in sample2)