node_position_path='./data')
model = Model(G.degree(),
size=representation_size,
table_size=5000000,
input_file=os.path.join(input_file, input_file + "_zachary"),
path_labels="./data")
model.node_color = node_color
# Sampling the random walks for context
if sampling_path:
log.info("sampling the paths")
walk_files = graph_utils.write_walks_to_disk(
G,
walks_filebase,
num_paths=number_walks,
path_length=walk_length,
alpha=0,
rand=random.Random(9999999999),
num_workers=num_workers)
else:
walk_files = [
walks_filebase + '.' + str(i) for i in range(number_walks)
if os.path.isfile(walks_filebase + '.' + str(i))
]
#Learning algorithm
node_learner = Node2Vec(workers=num_workers, negative=negative, lr=lr)
cont_learner = Context2Vec(window_size=window_size,
workers=num_workers,
negative=negative,
lr=lr)
ks = [2] # number of communities
walks_filebase = os.path.join(
'data', input_file) # where read/write the sampled path
# CONSTRUCT THE GRAPH
# G = graph_utils.load_matfile(os.path.join('./data', input_file, input_file + '.mat'), undirected=True)
G = graph_utils.load_csv_edges(os.path.join('./data', input_file,
input_file + '.csv'),
undirected=True)
# Sampling the random walks for context
log.info("sampling the paths")
walk_files = graph_utils.write_walks_to_disk(
G,
os.path.join(walks_filebase, "{}.walks".format(output_file)),
num_paths=number_walks,
path_length=walk_length,
alpha=0,
rand=random.Random(0),
num_workers=num_workers)
vertex_counts = graph_utils.count_textfiles(walk_files, num_workers)
model = Model(vertex_counts,
size=representation_size,
down_sampling=down_sampling,
table_size=100000000,
input_file=os.path.join(input_file, input_file),
path_labels="./data")
# Learning algorithm
node_learner = Node2Vec(workers=num_workers, negative=negative, lr=lr)
cont_learner = Context2Vec(window_size=window_size,
graph_name="karate",
node_position_file=True,
node_position_path='./data')
exmple_filebase = os.path.join(
"./data/",
output_file + ".exmple") # where read/write the sampled path
num_iter = G.number_of_nodes() * num_walks * walk_length
# Sampling the random walks for context
log.info("sampling the paths")
examples_files = graph_utils.write_walks_to_disk(
G,
exmple_filebase,
windows_size=window_size,
num_paths=num_walks,
path_length=walk_length,
alpha=0,
rand=random.Random(9999999999),
num_workers=num_workers)
edges = np.array(G.edges())
edges = np.concatenate((edges, np.fliplr(edges)))
io_utils.save_embedding(model.get_node_embedding(),
"pytorch_embedding_random",
path="./data")
# pre-training phase
learn_second(o2_loss,
lr,
model,
