def visualize(sess, config, args):
from tensorflow.python import debug as tf_debug
from kgcn.visualization import cal_feature_IG
# 入力は１分子づつ
batch_size = 1
# 入力データから、全データの情報, 学習用データの情報, 検証用データの情報, および
# グラフに関する情報を順に取得する
all_data, info = load_data(config,
filename=config["dataset"],
prohibit_shuffle=True)
#all_data.labels = tf.one_hot(tf.cast(tf.squeeze(all_data.labels), tf.int32), depth=2)
model = importlib.import_module(config["model.py"])
placeholders = model.build_placeholders(info,
config,
batch_size=batch_size)
try:
_model, prediction, _, _, _ = model.build_model(
placeholders,
info,
config,
batch_size=batch_size,
feed_embedded_layer=True)
except:
_model, prediction, _, _, _ = model.build_model(placeholders,
info,
config,
batch_size=batch_size)
#--- セッションの初期化
saver = tf.train.Saver()
print("[LOAD]", config["load_model"])
saver.restore(sess, config["load_model"])
#--- integrated gradientsの計算
cal_feature_IG(sess,
all_data,
placeholders,
info,
prediction,
args.ig_modal_target,
args.ig_label_target,
logger=tf.logging,
model=_model)
def visualize(sess, config, args):
from kgcn.visualization import cal_feature_IG, cal_feature_IG_for_kg
# input a molecule at a time
batch_size = 1
dataset_filename = config["dataset"]
if "dataset_test" in config:
dataset_filename = config["dataset_test"]
all_data, info = load_data(config,
filename=dataset_filename,
prohibit_shuffle=True)
model = CoreModel(sess, config, info)
load_model_py(model,
config["model.py"],
is_train=False,
feed_embedded_layer=True,
batch_size=batch_size)
placeholders = model.placeholders
restore_ckpt(sess, config['load_model'])
# calculate integrated gradients
if config['visualize_type'] == 'graph':
cal_feature_IG(sess,
all_data,
placeholders,
info,
config,
model.prediction,
args.ig_modal_target,
args.ig_label_target,
logger=tf.logging,
model=model.nn,
args=args)
else:
cal_feature_IG_for_kg(sess,
all_data,
placeholders,
info,
config,
model.prediction,
logger=tf.logging,
model=model.nn,
args=args)
def visualize(sess, config, args):
from tensorflow.python import debug as tf_debug
from kgcn.visualization import cal_feature_IG, cal_feature_IG_for_kg
# 入力は１分子づつ
batch_size = 1
# 入力データから、全データの情報, 学習用データの情報, 検証用データの情報, および
# グラフに関する情報を順に取得する
all_data, info = load_data(config,
filename=config["dataset"],
prohibit_shuffle=True)
model = importlib.import_module(config["model.py"])
try:
# emmbedingレイヤを使っているモデルの可視化。IGはemmbedingレイヤの出力を対象にして計算される。
placeholders = model.build_placeholders(info,
config,
batch_size=batch_size,
feed_embedded_layer=True)
except:
placeholders = model.build_placeholders(info,
config,
batch_size=batch_size)
try:
# emmbedingレイヤを使っているモデルの可視化。IGはemmbedingレイヤの出力を対象にして計算される。
_model, prediction, _, _, _ = model.build_model(
placeholders,
info,
config,
batch_size=batch_size,
feed_embedded_layer=True)
except:
_model, prediction, _, _, _ = model.build_model(placeholders,
info,
config,
batch_size=batch_size)
#--- セッションの初期化
saver = tf.train.Saver()
#tf.compat.v1.logging.info("[LOAD]", config["load_model"])
tf.logging.info("[LOAD]", config["load_model"])
saver.restore(sess, config["load_model"])
#--- integrated gradientsの計算
if config['visualize_type'] == 'graph':
cal_feature_IG(sess,
all_data,
placeholders,
info,
prediction,
args.ig_modal_target,
args.ig_label_target,
logger=tf.logging,
model=_model)
#logger=tf.compat.v1.logging, model=_model)
else:
cal_feature_IG_for_kg(sess,
all_data,
placeholders,
info,
config,
prediction,
logger=tf.logging,
model=_model)