Code for medium article about RL on Snake

AIP0-training-loop.py

jList = []
rList = []
for i in range(num_episodes):
    s = env.clear_board()
    rAll = 0
    d = False
    j = 0
    while j < game_max_length:
      # Game step...
    if i % save_interval == 0 and i > 0:
        print('Done {}/{} {:.2f}% Last game length: {}, Average game length: {}, Average length in last {}: {}, average R in last {}: {}'.format(i, num_episodes, i/num_episodes*100,j, sum(jList)/i, save_interval, sum(jList[-save_interval:])/save_interval, save_interval, sum(rList[-save_interval:])/save_interval))
        rlModel.saveVariables()
    jList.append(j)
    rList.append(rAll)
print("Average R: " + str(sum(rList)/num_episodes))
print("Average lenght: " + str(sum(jList)/num_episodes))