def get_ball():
return make_ball_gif(n_frames=n_time_steps,
f_height=f_height,
f_width=f_width,
ball_size=ball_size,
max_1=True,
vector_output=True)
# combined.append(x_reconstruct[i])
# combined.append(x_reconstruct[i+1])
# combined.append(x_reconstruct[i+2])
# kargs = { 'duration': .8 }
# imageio.mimsave(home+"/Downloads/comb_gif.gif", combined, 'GIF', **kargs)
#PREDICT
batch = []
while len(batch) != batch_size:
frame=make_ball_gif(n_frames=1, f_height=f_height, f_width=f_width, ball_size=ball_size, max_1=True, vector_output=True)
batch.append(frame)
batch = np.array(batch)
batch = np.reshape(batch, [batch_size, 1 , f_height*f_width])
print batch.shape
# predicted = vae.call_predict_next(batch)
# print predicted.shape
# together = [batch[0][0], predicted[0]]
# together = np.array(together)
# together = np.reshape(together, [2, f_height,f_width, 1])
# for i in range(len(together)):
# together[i] = together[i] * (255. / np.max(together[i]))
from __future__ import absolute_import # from __future__ import print_function import matplotlib.pyplot as plt import autograd.numpy as np import autograd.numpy.random as npr import autograd.scipy.stats.norm as norm # from black_box_svi import black_box_variational_inference from autograd.optimizers import adam from ball_sequence import make_ball_gif sequence, action_list = (make_ball_gif()) print sequence.shape
