if d :
bef_o = env.reset()
else:
bef_o = aft_o
if not args.vm :
env.render()
my_tqdm.close()
next_save = player.save_model()
if not args.load:
save_dir = player.save_dir
else:
save_dir, _ = os.path.split(args.load)
next_dir = os.path.join(save_dir,str(next_save))
score = player.evaluate(gym.make(ENVIRONMENT), vid_type)
print('eval_score:{0}'.format(score))
print('{0}steps took {1} sec'.format(total_steps,time.time()-st))
total_loop -= 1
if total_loop <= 0 :
sys.exit()
else :
next_args = []
next_args.append('python')
next_args.append(__file__)
next_args.append('-v')
next_args.append('-l')
next_args.append(next_dir)
next_args.append('--step')
next_args.append(str(total_steps))
for step in trange(total_steps, ncols=50):
action = player.act(o)
o, r, d, i = env.step(action)
player.step(o, r, d, i)
if d:
o = env.reset()
if not args.vm:
env.render()
next_save = player.save_model()
if not args.load:
save_dir = player.save_dir
else:
save_dir, _ = os.path.split(args.load)
next_dir = os.path.join(save_dir, str(next_save))
score = player.evaluate(gym.make('mouseCl-v0'), vid_type)
print('eval_score:{0}'.format(score))
print('{0}steps took {1} sec'.format(total_steps, time.time() - st))
total_loop -= 1
if total_loop <= 0:
sys.exit()
else:
next_args = []
next_args.append('python')
next_args.append(__file__)
next_args.append('-v')
next_args.append('-l')
next_args.append(next_dir)
next_args.append('--step')
next_args.append(str(total_steps))