def train_gpmcc_chain(num_points=100, num_states=1, num_transitions = 100, save_out=True):
data = simulate_chain(num_points, rng = np.random.RandomState(0))
start_time = time.time()
engine = Engine(data, ['normal','normal','normal'], num_states=num_states, initialize=1)
engine.transition(N=num_transitions, do_progress=True, do_plot=False)
train_time = time.time() - start_time
engine.num_points = num_points
engine.num_transitions = num_transitions
engine.data = data
if save_out:
file_engine = file('resources/eng_chain_pnts%d_stats%d_trans%d.pkl' % (num_points, num_states, num_transitions),'wb')
file_data_time =file('resources/data_time_chain_pnts%d_stats%d_trans%d.pkl' % (num_points, num_states, num_transitions),'wb')
engine.to_pickle(file_engine)
pickle.dump([data, train_time],file_data_time)
return engine, data, train_time
def load_gpmcc_chain(num_points=100, num_states=1, num_transitions = 100):
file_engine = file('resources/eng_chain_pnts%d_stats%d_trans%d.pkl' % (num_points, num_states, num_transitions),'rb')
file_data_time =file('resources/data_time_chain_pnts%d_stats%d_trans%d.pkl' % (num_points, num_states, num_transitions),'rb')
engine = Engine.from_pickle(file_engine)
data, train_time = pickle.load(file_data_time)
return engine, data, train_time