-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_DecTag_NFM.py
131 lines (114 loc) · 5.79 KB
/
test_DecTag_NFM.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
import os
import time
import argparse
import numpy as np
import random
import pandas as pd
import torch
import torch.nn as nn
import torch.optim as optim
import torch.utils.data as data
import torch.backends.cudnn as cudnn
import model
import data_util
import evaluate_NFM
parser = argparse.ArgumentParser()
parser.add_argument("--dataset",
type=str,
help="Dataset")
parser.add_argument("--splits",
type=str,
default='1,2,3,4,5',
help="split index")
parser.add_argument("--gpu",
type=str,
default="0",
help="gpu card ID")
parser.add_argument("--seed",
type=int,
default=1,
help="random seed")
args = parser.parse_args()
args.splits = args.splits.split(',')
os.environ["CUDA_VISIBLE_DEVICES"] = args.gpu
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
random_seed = args.seed
torch.manual_seed(random_seed)
torch.cuda.manual_seed(random_seed)
np.random.seed(random_seed)
random.seed(random_seed)
torch.cuda.manual_seed_all(random_seed)
torch.backends.cudnn.benchmark = False
torch.backends.cudnn.deterministic = True
Recall = []
NDCG = []
topks = [10, 20]
for k in topks:
Recall.append([])
NDCG.append([])
data_set = args.dataset
print("Start testing Dataset {} (splits {}) for DecTag-NFM".format(args.dataset, args.splits))
for i in range(len(args.splits)):
######## RESULT DAFAFRAME ########
df_columns = ['split_index', 'model_state_dict_path']
for k in topks:
df_columns.append('recall@{}'.format(k))
for k in topks:
df_columns.append('ndcg@{}'.format(k))
result_df = pd.DataFrame(columns=df_columns, index=[0])
result_df['split_index'][0] = args.splits[i]
######## PREPARE DATASET ########
split_index = args.splits[i]
start_time = time.time()
item_features_path = 'data_split/{}/video_infos.npy'.format(data_set)
pair_list_path = 'data_split/{}/pair_list.npy'.format(data_set)
user_features_path = 'data_split/{}/user_features.npy'.format(data_set)
video_features_path = 'data_split/{}/video_features.npy'.format(data_set)
tag_num_path = 'data_split/{}/split_{}/tag_num.npy'.format(data_set, split_index)
item_features = np.load(item_features_path, allow_pickle=True)[()]
pair_list = np.load(pair_list_path, allow_pickle=True)[()]
user_features = np.load(user_features_path, allow_pickle=True)[()]
tag_num = np.load(tag_num_path, allow_pickle=True)[()]
videohiddenid_itemid_map, itemid_videohiddenid_map, channelurl_userid_map, _, _ = data_util.generate_map_dict(item_features, user_features, pair_list)
train_pair_list_path = 'data_split/{}/split_{}/train_pair_list.npy'.format(data_set, split_index)
valid_pair_list_path = 'data_split/{}/split_{}/valid_pair_list.npy'.format(data_set, split_index)
test_pair_list_path = 'data_split/{}/split_{}/test_pair_list.npy'.format(data_set, split_index)
item_confounder_path = 'data_split/{}/split_{}/item_confounder.npy'.format(data_set, split_index)
confounders_path = 'data_split/{}/split_{}/confounders.npy'.format(data_set, split_index)
train_pair_list = np.load(train_pair_list_path, allow_pickle=True)[()]
valid_pair_list = np.load(valid_pair_list_path, allow_pickle=True)[()]
test_pair_list = np.load(test_pair_list_path, allow_pickle=True)[()]
item_confounder = np.load(item_confounder_path, allow_pickle=True)[()]
confounders = np.load(confounders_path, allow_pickle=True)[()]
######## CREATE MODEL ########
pth_file = os.listdir("check_point/{}/NFM/split_{}".format(data_set, split_index))[0]
pre_train_path = os.path.join("check_point/{}/NFM/split_{}".format(data_set, split_index), pth_file)
result_df['model_state_dict_path'][0] = pre_train_path
_model = model.DecTag_NFM(tag_num=tag_num, confounders=confounders, num_layers=3, dropout=0.3, category_num=2).to(device)
_model.load_state_dict(torch.load(pre_train_path))
######## TESTING ########
_model.eval()
test_result = \
evaluate_NFM.test_Ranking(_model, video_features_path, itemid_videohiddenid_map, test_pair_list, train_pair_list, tag_num=tag_num, topks=topks, split_index=split_index)
print("----" * 18)
for j in range(len(topks)):
print("Split{} Testing Set: Recall@{}: {:.5f} NDCG@{}: {:.5f}".format(split_index, topks[j], test_result['recall'][j], topks[j], test_result['ndcg'][j]))
result_df['recall@{}'.format(topks[j])][0] = test_result['recall'][j]
result_df['ndcg@{}'.format(topks[j])][0] = test_result['ndcg'][j]
print("----" * 18)
######## SAVING RESULT ########
now_time = time.strftime("%Y-%m-%d-%H_%M_%S", time.localtime())
if os.path.exist('results') == False:
os.mkdir('results')
if os.path.exist('results/{}'.format(data_set)) == False:
os.mkdir('results/{}'.format(data_set))
if os.path.exist('results/{}/NFM/split_{}'.format(data_set, split_index)) == False:
os.mkdir('results/{}/NFM/split_{}'.format(data_set, split_index))
result_df.to_csv('results/{}/NFM/split_{}/test_result_{}.csv'.format(data_set, split_index, now_time), sep=',', header=True, index=False)
for j in range(len(topks)):
Recall[j].append(test_result['recall'][j])
NDCG[j].append(test_result['ndcg'][j])
for j in range(len(topks)):
print("Final(mean): Recall@{}: {:.5f} NDCG@{}: {:.5f}".format(topks[j], np.mean(np.array(Recall[j])), topks[j], np.mean(np.array(NDCG[j]))))
for j in range(len(topks)):
print("Final(std): Recall@{}: {:.5f} NDCG@{}: {:.5f}".format(topks[j], np.std(np.array(Recall[j])), topks[j], np.std(np.array(NDCG[j]))))