本文整理汇总了Python中src.graph_utils.heuristic_fn_vec方法的典型用法代码示例。如果您正苦于以下问题：Python graph_utils.heuristic_fn_vec方法的具体用法？Python graph_utils.heuristic_fn_vec怎么用？Python graph_utils.heuristic_fn_vec使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类src.graph_utils的用法示例。
在下文中一共展示了graph_utils.heuristic_fn_vec方法的2个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: _compute_hardness
# 需要导入模块: from src import graph_utils [as 别名]
# 或者: from src.graph_utils import heuristic_fn_vec [as 别名]
def _compute_hardness():
# Load the stanford data to compute the hardness.
if FLAGS.type == '':
args = sna.get_args_for_config(FLAGS.config_name+'+bench_'+FLAGS.imset)
else:
args = sna.get_args_for_config(FLAGS.type+'.'+FLAGS.config_name+'+bench_'+FLAGS.imset)
args.navtask.logdir = None
R = lambda: nav_env.get_multiplexer_class(args.navtask, 0)
R = R()
rng_data = [np.random.RandomState(0), np.random.RandomState(0)]
# Sample a room.
h_dists = []
gt_dists = []
for i in range(250):
e = R.sample_env(rng_data)
nodes = e.task.nodes
# Initialize the agent.
init_env_state = e.reset(rng_data)
gt_dist_to_goal = [e.episode.dist_to_goal[0][j][s]
for j, s in enumerate(e.episode.start_node_ids)]
for j in range(args.navtask.task_params.batch_size):
start_node_id = e.episode.start_node_ids[j]
end_node_id =e.episode.goal_node_ids[0][j]
h_dist = graph_utils.heuristic_fn_vec(
nodes[[start_node_id],:], nodes[[end_node_id], :],
n_ori=args.navtask.task_params.n_ori,
step_size=args.navtask.task_params.step_size)[0][0]
gt_dist = e.episode.dist_to_goal[0][j][start_node_id]
h_dists.append(h_dist)
gt_dists.append(gt_dist)
h_dists = np.array(h_dists)
gt_dists = np.array(gt_dists)
e = R.sample_env([np.random.RandomState(0), np.random.RandomState(0)])
input = e.get_common_data()
orig_maps = input['orig_maps'][0,0,:,:,0]
return h_dists, gt_dists, orig_maps 示例2: _compute_hardness
# 需要导入模块: from src import graph_utils [as 别名]
# 或者: from src.graph_utils import heuristic_fn_vec [as 别名]
def _compute_hardness():
# Load the stanford data to compute the hardness.
if FLAGS.type == '':
args = sna.get_args_for_config(FLAGS.config_name+'+bench_'+FLAGS.imset)
else:
args = sna.get_args_for_config(FLAGS.type+'.'+FLAGS.config_name+'+bench_'+FLAGS.imset)
args.navtask.logdir = None
R = lambda: nav_env.get_multiplexer_class(args.navtask, 0)
R = R()
rng_data = [np.random.RandomState(0), np.random.RandomState(0)]
# Sample a room.
h_dists = []
gt_dists = []
for i in range(250):
e = R.sample_env(rng_data)
nodes = e.task.nodes
# Initialize the agent.
init_env_state = e.reset(rng_data)
gt_dist_to_goal = [e.episode.dist_to_goal[0][j][s]
for j, s in enumerate(e.episode.start_node_ids)]
for j in range(args.navtask.task_params.batch_size):
start_node_id = e.episode.start_node_ids[j]
end_node_id =e.episode.goal_node_ids[0][j]
h_dist = graph_utils.heuristic_fn_vec(
nodes[[start_node_id],:], nodes[[end_node_id], :],
n_ori=args.navtask.task_params.n_ori,
step_size=args.navtask.task_params.step_size)[0][0]
gt_dist = e.episode.dist_to_goal[0][j][start_node_id]
h_dists.append(h_dist)
gt_dists.append(gt_dist)
h_dists = np.array(h_dists)
gt_dists = np.array(gt_dists)
e = R.sample_env([np.random.RandomState(0), np.random.RandomState(0)])
input = e.get_common_data()
orig_maps = input['orig_maps'][0,0,:,:,0]
return h_dists, gt_dists, orig_maps