def train_cv(sess, graph, config):
all_data, info = load_data(
config, filename=config["dataset"],
prohibit_shuffle=True) # shuffle is done by KFold
model = CoreModel(sess, config, info)
load_model_py(model, config["model.py"])
# Training
if config["stratified_kfold"]:
print("[INFO] use stratified K-fold")
kf = StratifiedKFold(n_splits=config["k-fold_num"],
shuffle=config["shuffle_data"],
random_state=123)
else:
kf = KFold(n_splits=config["k-fold_num"],
shuffle=config["shuffle_data"],
random_state=123)
kf_count = 1
fold_data_list = []
output_data_list = []
if all_data["labels"] is not None:
split_base = all_data["labels"]
else:
split_base = all_data["label_list"][0]
if config["stratified_kfold"]:
split_base = np.argmax(split_base, axis=1)
score_metrics = []
if config["task"] == "regression":
metric_name = "mse"
elif config["task"] == "regression_gmfe":
metric_name = "gmfe"
else:
metric_name = "accuracy"
split_data_generator = kf.split(
split_base,
split_base) if config["stratified_kfold"] else kf.split(split_base)
for train_valid_list, test_list in split_data_generator:
print(f"starting fold: {kf_count}")
train_valid_data, test_data = split_data(
all_data,
indices_for_train_data=train_valid_list,
indices_for_valid_data=test_list)
train_data, valid_data = split_data(
train_valid_data, valid_data_rate=config["validation_data_rate"])
# Training
print(train_valid_list)
print(test_list)
start_t = time.time()
model.fit(train_data, valid_data, k_fold_num=kf_count)
train_time = time.time() - start_t
print(f"training time: {train_time}[sec]")
# Test
print("== valid data ==")
start_t = time.time()
valid_cost, valid_metrics, prediction_data = model.pred_and_eval(
valid_data)
infer_time = time.time() - start_t
print(f"final cost = {valid_cost}\n"
f"{metric_name} = {valid_metrics[metric_name]}\n"
f"infer time: {infer_time}[sec]\n")
print("== test data ==")
start_t = time.time()
test_cost, test_metrics, prediction_data = model.pred_and_eval(
test_data)
infer_time = time.time() - start_t
print(f"final cost = {test_cost}\n"
f"{metric_name} = {test_metrics[metric_name]}\n")
score_metrics.append(test_metrics[metric_name])
print(f"infer time: {infer_time}[sec]")
if config["export_model"]:
try:
name, ext = os.path.splitext(config["export_model"])
filename = name + "." + str(kf_count) + ext
print(f"[SAVE] {filename}")
graph_def = graph_util.convert_variables_to_constants(
sess, graph.as_graph_def(), ['output'])
tf.train.write_graph(graph_def, '.', filename, as_text=False)
except:
print('[ERROR] output has been not found')
if "save_edge_result_cv" in config:
output_data = model.output(test_data)
output_data_list.append(output_data)
# save fold data
fold_data = dotdict({})
fold_data.prediction_data = prediction_data
if all_data["labels"] is not None:
fold_data.test_labels = test_data.labels
else:
fold_data.test_labels = test_data.label_list
fold_data.test_data_idx = test_list
if config["task"] == "regression":
fold_data.training_mse = [
el["training_mse"] for el in model.training_metrics_list
]
fold_data.validation_mse = [
el["validation_mse"] for el in model.validation_metrics_list
]
elif config["task"] == "regression_gmfe":
fold_data.training_mse = [
el["training_gmfe"] for el in model.training_metrics_list
]
fold_data.validation_mse = [
el["validation_gmfe"] for el in model.validation_metrics_list
]
else:
fold_data.training_acc = [
el["training_accuracy"] for el in model.training_metrics_list
]
fold_data.validation_acc = [
el["validation_accuracy"]
for el in model.validation_metrics_list
]
fold_data.test_acc = test_metrics[metric_name]
fold_data.training_cost = model.training_cost_list
fold_data.validation_cost = model.validation_cost_list
fold_data.test_cost = test_cost
fold_data.train_time = train_time
fold_data.infer_time = infer_time
fold_data_list.append(fold_data)
kf_count += 1
print(f"cv {metric_name}(mean) = {np.mean(score_metrics)}\n"
f"cv {metric_name}(std.) = {np.std(score_metrics)}\n")
if "save_info_cv" in config and config["save_info_cv"] is not None:
save_path = config["save_info_cv"]
os.makedirs(os.path.dirname(save_path), exist_ok=True)
print(f"[SAVE] {save_path}")
_, ext = os.path.splitext(save_path)
if ext == ".json":
with open(save_path, "w") as fp:
json.dump(fold_data_list, fp, indent=4, cls=NumPyArangeEncoder)
else:
joblib.dump(fold_data_list, save_path, compress=True)
#
if "save_edge_result_cv" in config and config[
"save_edge_result_cv"] is not None:
result_cv = []
for j, fold_data in enumerate(fold_data_list):
pred_score = np.array(fold_data.prediction_data)
true_label = np.array(fold_data.test_labels)
test_idx = fold_data.test_data_idx
score_list = []
for pair in true_label[0]:
i1, _, j1, i2, _, j2 = pair
s1 = pred_score[0, i1, j1]
s2 = pred_score[0, i2, j2]
score_list.append([s1, s2])
fold = {}
fold["output"] = output_data_list[j][0]
fold["score"] = np.array(score_list)
fold["test_data_idx"] = test_idx
result_cv.append(fold)
save_path = config["save_edge_result_cv"]
os.makedirs(os.path.dirname(save_path), exist_ok=True)
print(f"[SAVE] {save_path}")
_, ext = os.path.splitext(save_path)
if ext == ".json":
with open(save_path, "w") as fp:
json.dump(result_cv, fp, indent=4, cls=NumPyArangeEncoder)
else:
joblib.dump(result_cv, save_path, compress=True)
#
if "save_result_cv" in config and config["save_result_cv"] is not None:
result_cv = []
for j, fold_data in enumerate(fold_data_list):
v = compute_metrics(config, info, fold_data.prediction_data,
fold_data.test_labels)
result_cv.append(v)
save_path = config["save_result_cv"]
print(f"[SAVE] {save_path}")
with open(save_path, "w") as fp:
json.dump(result_cv, fp, indent=4, cls=NumPyArangeEncoder)
#
for i, fold_data in enumerate(fold_data_list):
prefix = "fold" + str(i) + "_"
result_path = config["plot_path"]
os.makedirs(result_path, exist_ok=True)
if config["make_plot"]:
if config["task"] == "regression":
make_cost_acc_plot(fold_data.training_cost,
fold_data.validation_cost,
fold_data.training_mse,
fold_data.validation_mse,
result_path,
prefix=prefix)
pred_score = np.array(fold_data.prediction_data)
plot_r2(config,
fold_data.test_labels,
pred_score,
prefix=prefix)
elif config["task"] == "regression_gmfe":
make_cost_acc_plot(fold_data.training_cost,
fold_data.validation_cost,
fold_data.training_mse,
fold_data.validation_mse,
result_path,
prefix=prefix)
pred_score = np.array(fold_data.prediction_data)
plot_r2(config,
fold_data.test_labels,
pred_score,
prefix=prefix)
elif config["task"] == "link_prediction":
make_cost_acc_plot(fold_data.training_cost,
fold_data.validation_cost,
fold_data.training_acc,
fold_data.validation_acc,
result_path,
prefix=prefix)
else:
make_cost_acc_plot(fold_data.training_cost,
fold_data.validation_cost,
fold_data.training_acc,
fold_data.validation_acc,
result_path,
prefix=prefix)
pred_score = np.array(fold_data.prediction_data)
plot_auc(config,
fold_data.test_labels,
pred_score,
prefix=prefix)
def train_cv(sess, graph, config):
from sklearn.model_selection import KFold, StratifiedKFold
from kgcn.make_plots import make_auc_plot, make_cost_acc_plot
import sklearn
from sklearn.metrics import roc_curve, auc, accuracy_score, precision_recall_fscore_support
from scipy import interp
batch_size = config["batch_size"]
learning_rate = config["learning_rate"]
all_data, info = load_data(
config, filename=config["dataset"],
prohibit_shuffle=True) # shuffle is done by KFold
model = CoreModel(sess, config, info)
load_model_py(model, config["model.py"])
# Training
if config["stratified_kfold"]:
print("[INFO] use stratified K-fold")
kf = StratifiedKFold(n_splits=config["k-fold_num"],
shuffle=config["shuffle_data"],
random_state=123)
else:
kf = KFold(n_splits=config["k-fold_num"],
shuffle=config["shuffle_data"],
random_state=123)
kf_count = 1
fold_data_list = []
output_data_list = []
if all_data["labels"] is not None:
split_base = all_data["labels"]
else:
split_base = all_data["label_list"][0]
if config["stratified_kfold"]:
split_base = np.argmax(split_base, axis=1)
score_metrics = []
if config["task"] == "regression":
metric_name = "mse"
elif config["task"] == "regression_gmfe":
metric_name = "gmfe"
else:
metric_name = "accuracy"
split_data_generator = kf.split(
split_base,
split_base) if config["stratified_kfold"] else kf.split(split_base)
for train_valid_list, test_list in split_data_generator:
print("starting fold:{0}".format(kf_count))
train_valid_data, test_data = split_data(
all_data,
indices_for_train_data=train_valid_list,
indices_for_valid_data=test_list)
train_data, valid_data = split_data(
train_valid_data, valid_data_rate=config["validation_data_rate"])
# Training
print(train_valid_list)
print(test_list)
start_t = time.time()
model.fit(train_data, valid_data, k_fold_num=kf_count)
train_time = time.time() - start_t
print("traing time:{0}".format(train_time) + "[sec]")
# Test
print("== valid data ==")
start_t = time.time()
valid_cost, valid_metrics, prediction_data = model.pred_and_eval(
valid_data)
infer_time = time.time() - start_t
print("final cost =", valid_cost)
print("%s =%f" % (metric_name, valid_metrics[metric_name]))
print("infer time:{0}".format(infer_time) + "[sec]")
print("== test data ==")
start_t = time.time()
test_cost, test_metrics, prediction_data = model.pred_and_eval(
test_data)
infer_time = time.time() - start_t
print("final cost =", test_cost)
print("%s =%f" % (metric_name, test_metrics[metric_name]))
score_metrics.append(test_metrics[metric_name])
print("infer time:{0}".format(infer_time) + "[sec]")
if config["export_model"]:
try:
name, ext = os.path.splitext(config["export_model"])
filename = name + "." + str(kf_count) + ext
print("[SAVE]", filename)
graph_def = graph_util.convert_variables_to_constants(
sess, graph.as_graph_def(), ['output'])
tf.train.write_graph(graph_def, '.', filename, as_text=False)
except:
print('[ERROR] output has been not found')
if "save_edge_result_cv" in config:
output_data = model.output(test_data)
output_data_list.append(output_data)
# save fold data
fold_data = dotdict({})
fold_data.prediction_data = prediction_data
if all_data["labels"] is not None:
fold_data.test_labels = test_data.labels
else:
fold_data.test_labels = test_data.label_list
fold_data.test_data_idx = test_list
if config["task"] == "regression":
fold_data.training_mse = [
el["training_mse"] for el in model.training_metrics_list
]
fold_data.validation_mse = [
el["validation_mse"] for el in model.validation_metrics_list
]
elif config["task"] == "regression_gmfe":
fold_data.training_mse = [
el["training_gmfe"] for el in model.training_metrics_list
]
fold_data.validation_mse = [
el["validation_gmfe"] for el in model.validation_metrics_list
]
else:
fold_data.training_acc = [
el["training_accuracy"] for el in model.training_metrics_list
]
fold_data.validation_acc = [
el["validation_accuracy"]
for el in model.validation_metrics_list
]
fold_data.test_acc = test_metrics[metric_name]
fold_data.training_cost = model.training_cost_list
fold_data.validation_cost = model.validation_cost_list
fold_data.test_cost = test_cost
fold_data.train_time = train_time
fold_data.infer_time = infer_time
fold_data_list.append(fold_data)
kf_count += 1
print("cv %s(mean) =%f" % (metric_name, np.mean(score_metrics)))
print("cv %s(std.) =%f" % (metric_name, np.std(score_metrics)))
if "save_info_cv" in config and config["save_info_cv"] is not None:
save_path = config["save_info_cv"]
os.makedirs(os.path.dirname(save_path), exist_ok=True)
print("[SAVE] ", save_path)
_, ext = os.path.splitext(save_path)
if ext == ".json":
fp = open(save_path, "w")
json.dump(fold_data_list, fp, indent=4, cls=NumPyArangeEncoder)
else:
joblib.dump(fold_data_list, save_path, compress=True)
##
if "save_edge_result_cv" in config and config[
"save_edge_result_cv"] is not None:
result_cv = []
for j, fold_data in enumerate(fold_data_list):
pred_score = np.array(fold_data.prediction_data)
true_label = np.array(fold_data.test_labels)
test_idx = fold_data.test_data_idx
score_list = []
for pair in true_label[0]:
i1, _, j1, i2, _, j2 = pair
s1 = pred_score[0, i1, j1]
s2 = pred_score[0, i2, j2]
score_list.append([s1, s2])
fold = {}
fold["output"] = output_data_list[j][0]
fold["score"] = np.array(score_list)
fold["test_data_idx"] = test_idx
result_cv.append(fold)
save_path = config["save_edge_result_cv"]
os.makedirs(os.path.dirname(save_path), exist_ok=True)
print("[SAVE] ", save_path)
_, ext = os.path.splitext(save_path)
if ext == ".json":
fp = open(save_path, "w")
json.dump(result_cv, fp, indent=4, cls=NumPyArangeEncoder)
else:
joblib.dump(result_cv, save_path, compress=True)
#
#
if "save_result_cv" in config and config["save_result_cv"] is not None:
result_cv = []
for j, fold_data in enumerate(fold_data_list):
pred_score = np.array(fold_data.prediction_data)
if len(pred_score.shape) == 3: # multi-label-multi-task
# #data x # task x #class
# => this program supports only 2 labels
pred_score = pred_score[:, :, 1]
true_label = np.array(fold_data.test_labels)
# #data x # task x #class
if len(pred_score.shape) == 1:
pred_score = pred_score[:, np.newaxis]
if len(true_label.shape) == 1:
true_label = true_label[:, np.newaxis]
v = []
for i in range(info.label_dim):
el = {}
if config["task"] == "regression":
el["r2"] = sklearn.metrics.r2_score(
true_label[:, i], pred_score[:, i])
el["mse"] = sklearn.metrics.mean_squared_error(
true_label[:, i], pred_score[:, i])
elif config["task"] == "regression_gmfe":
el["gmfe"] = np.exp(
np.mean(np.log(true_label[:, i] / pred_score[:, i])))
else:
pred = np.zeros(pred_score.shape)
pred[pred_score > 0.5] = 1
fpr, tpr, _ = roc_curve(true_label[:, i],
pred_score[:, i],
pos_label=1)
roc_auc = auc(fpr, tpr)
acc = accuracy_score(true_label[:, i], pred[:, i])
scores = precision_recall_fscore_support(true_label[:, i],
pred[:, i],
average='binary')
el["auc"] = roc_auc
el["acc"] = acc
el["pre"] = scores[0]
el["rec"] = scores[1]
el["f"] = scores[2]
el["sup"] = scores[3]
v.append(el)
result_cv.append(v)
save_path = config["save_result_cv"]
print("[SAVE] ", save_path)
fp = open(save_path, "w")
json.dump(result_cv, fp, indent=4, cls=NumPyArangeEncoder)
#
for i, fold_data in enumerate(fold_data_list):
prefix = "fold" + str(i) + "_"
result_path = config["plot_path"]
os.makedirs(result_path, exist_ok=True)
if config["make_plot"]:
if config["task"] == "regression":
# plot cost
make_cost_acc_plot(fold_data.training_cost,
fold_data.validation_cost,
fold_data.training_mse,
fold_data.validation_mse,
result_path + prefix)
pred_score = np.array(fold_data.prediction_data)
plot_r2(config,
fold_data.test_labels,
pred_score,
prefix=prefix)
elif config["task"] == "regression_gmfe":
# plot cost
make_cost_acc_plot(fold_data.training_cost,
fold_data.validation_cost,
fold_data.training_mse,
fold_data.validation_mse,
result_path + prefix)
pred_score = np.array(fold_data.prediction_data)
plot_r2(config,
fold_data.test_labels,
pred_score,
prefix=prefix)
else:
# plot cost
make_cost_acc_plot(fold_data.training_cost,
fold_data.validation_cost,
fold_data.training_acc,
fold_data.validation_acc,
result_path + prefix)
# plot AUC
pred_score = np.array(fold_data.prediction_data)
plot_auc(config,
fold_data.test_labels,
pred_score,
prefix=prefix)