def run_a_sample(checkpoint):
newnn = nn()
newnn.load(checkpoint)
spc = SelfPlay(newnn)
states, heads, scores, ids, moves = spc.playGames(0, 0, 1)
fig = plt.figure()
ax = plt.axes(xlim=(-0.5, 3.5), ylim=(-0.5, 3.5))
ax.axes.get_yaxis().set_visible(False)
ax.axes.get_xaxis().set_visible(False)
ims = []
for i in range(len(states)):
im = plt.imshow(states[i], animated=True)
title = ax.text(
0.5,
1.05,
"{}".format(i),
size=plt.rcParams["axes.titlesize"],
ha="center",
transform=ax.transAxes,
)
ims.append([im, title])
ani = animation.ArtistAnimation(fig, ims, interval=200, blit=False)
plt.show()
def __init__(self):
self.nn = nn()
self.action_space_high = 18
self.action_space_low = -18
self.target = [10/57,0,0]
self.initial = [0,0,0]
self.target_error = 0.1*sum([abs(x-y) for x,y in zip(self.target,self.initial)])
def gen_compare():
states = []
generation = [0, 100, 200, 383]
for i in range(4):
newnn = nn()
newnn.load('saves/generation_{}.ckpt'.format(generation[i]))
spc = SelfPlay(newnn)
if i < 3:
state, head, score, id, move = spc.playGames(0, 0, 1)
else:
state, head, score, id, move = spc.playGames(0, 0, 50)
state = find_best(state, score, id)
states.append(state)
print('done with generation {}'.format(i))
gamelength = []
for i in range(4):
gamelength.append(len(states[i]))
for i in range(np.max(gamelength)):
fig = plt.figure()
for j in range(4):
ax = plt.subplot(2, 2, 1 + j)
plt.xlim([-0.5, 3.5])
plt.ylim([-0.5, 3.5])
ax.axes.get_yaxis().set_visible(False)
ax.axes.get_xaxis().set_visible(False)
if i < gamelength[j]:
ax.imshow(states[j][i])
plt.title('Generation {}, move {}'.format(generation[j], i))
else:
ax.imshow(states[j][-1])
plt.title('Generation {}, move {}'.format(
generation[j], gamelength[j]))
fig.savefig('compare/compare{}.png'.format(i))
plt.close()