def main():
config = get_config_from_args()
visualizer = Visualizer(config)
if config.dir:
visualizer.replay_dir()
else:
visualizer.replay_file()
import numpy as np
np.random.seed(1) # NumPy
import random
random.seed(2) # Python
from tensorflow import set_random_seed
set_random_seed(3) # Tensorflow
from causality_detection.feed_forward import Evaluation
from visualization.visualizer import Visualizer
if __name__ == '__main__':
evaluation = Evaluation()
visualizer = Visualizer()
experiment_key = 'cnet_wiki_exp_0'
settings = {
'dataset_file': 'causal_pairs_dataset_1000.csv',
'result_file': 'results.json',
'embedding_model_file': 'files/GoogleNews-vectors-negative300.bin',
'causal_net_file': 'causal_net_1m.pickle',
'n_pair': 1000,
'n_expand': 0,
'result_key': 'cnet_wiki_exp_0'
}
evaluation.run_experiment(settings=settings)
# settings['threshold'] = 10
# settings['result_key'] = 'luo_threshold_10'
# evaluation.run_experiment_on_luos_method(settings)
robot_parameter=robot_params,
sim_parameter=sim_params)
pilcotrac = PILCOTRAC(robot_params,
sim_params,
robot_state,
pathHandler,
motion_model,
feedforward,
SUBS=150)
# Wait to load visualizer
print("Example Knicklenkung PILCO")
print("\trobot_state = ", robot_state)
# Visualization
vis = Visualizer(robot_params, pathHandler, robot_state)
# sleep(3.0)
SUBS = 400
T = 30
T_sim = 2500
J = 10
with tf.Session() as sess:
max_yerror = 0.3
X1, Y1, k = pilcotrac.rollout(None,
max_yerror,
data_mean=None,
data_std=None,
lookahead=1.22,
from visualization.visualizer import Visualizer
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("-k", "--key", default=None, help="Experiment key")
parser.add_argument("-v",
"--visualize",
default='no',
help="Generate output charts")
args = parser.parse_args()
experiment_key = args.key
visualize = args.visualize
evaluation = Evaluation()
visualizer = Visualizer()
settings = {
'dataset_file': 'causal_pairs_dataset_1000.csv',
'result_file': 'results.json',
'embedding_model_file': 'files/GoogleNews-vectors-negative300.bin',
'causal_net_file': 'causal_net_1m.pickle',
'n_pair': 1000,
'n_expand': 0,
'result_key': 'cnet_wiki_exp_0'
}
if experiment_key == 'cnet_wiki_exp_0':
evaluation.run_experiment(settings=settings)
if experiment_key == 'cnet_wiki_exp_1':
def main():
if len(sys.argv) < 2:
raise ValueError('At least one file or folder needs to be provided')
results = Summary.loads(*sys.argv[1:])
vis = Visualizer(results)
f = plt.figure()
ax = f.add_subplot(231)
vis.plot_metric(lambda s: s.precision, name='Precision', ax=ax, legend=False, print_params=True)
ax = f.add_subplot(232)
vis.plot_metric(lambda s: s.tp + s.fn, name='Sampled Positives', ax=ax, legend=False)
def _f1(p, r):
f1 = np.zeros(shape=p.shape)
pr_sum = p + r
i = pr_sum > 0
f1[i] = 2 * p[i] * r[i] / pr_sum[i]
return f1
ax = f.add_subplot(233)
vis.plot_metric(lambda s: _f1(s.precision, s.recall), name='F1 score', ax=ax, legend=False)
ax = f.add_subplot(234)
vis.plot_metric(lambda s: s.recall, name='Recall', ax=ax, legend=False)
ax = f.add_subplot(235)
vis.plot_metric(lambda s: s.unique, name='Number of unique actions', ax=ax, legend=False)
ax = f.add_subplot(236)
vis.plot_metric(lambda s: s.unique_positive, name='Number of unique positive actions', ax=ax, legend=True)
plt.show()