Adaptive-Contrastive-Learning-For-Sequential-Recommendation/templates.py at master · Russole/Adaptive-Contrastive-Learning-For-Sequential-Recommendation · GitHub

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
from MCLRec_Part import AdaptModelHyper
def set_template(args):
    if args.template is None:
        return

    else:
        args.mode = 'train'
        #args.dataset_code = 'beauty'
        args.min_rating = 0
        args.min_uc = 5
        args.min_sc = 5
        args.split = 'leave_one_out'

        args.dataloader_code = 'bert'
        batch = 256
        if args.dataset_code=='beauty':
            args.bert_dropout = 0.3
            args.tau = 0.3
            seq_len = 15
            args.num_positive = 4


            args.decay_step = 100
        elif args.dataset_code=='toys':
            args.decay_step = 100
            args.bert_dropout = 0.3
            args.tau = 0.3
            args.num_positive = 4
            seq_len = 20
        elif args.dataset_code=='grocery':
            args.decay_step = 100
            args.bert_dropout = 0.3
            args.tau = 0.3
            args.num_positive = 4
            seq_len = 20
        elif args.dataset_code=='Automotive':
            args.decay_step = 100
            args.bert_dropout = 0.3
            args.tau = 0.3
            args.num_positive = 4
            seq_len = 20
        elif args.dataset_code=='Video_Games':
            args.decay_step = 100
            args.bert_dropout = 0.3
            args.tau = 0.3
            args.num_positive = 4
            seq_len = 20
        elif 'ml' in args.dataset_code:
            args.decay_step = 50
            seq_len = 40
            args.tau = 0.3
            args.num_positive = 8
            args.bert_dropout = 0.2
        else:
            args.decay_step = 50
            seq_len = 20
            args.bert_dropout = 0.1

        args.train_batch_size = batch
        args.val_batch_size = batch
        args.test_batch_size = batch

        #args.train_negative_sampler_code = 'random'

        args.train_negative_sample_size = seq_len

        #args.train_negative_sampling_seed = 56789
        #args.test_negative_sampler_code = 'random'
        #args.test_negative_sample_size = seq_len
        #args.test_negative_sampling_seed = 98765

        args.trainer_code = 'bert'
        #args.device = 'cuda:2'
        args.num_gpu = 1
        args.optimizer = 'Adam'
        args.lr = 0.001
        args.enable_lr_schedule = True

        args.gamma = 1.0

        args.num_epochs = 250
        args.metric_ks = [1, 5, 10, 20]
        args.best_metric = 'NDCG@10'

        args.model_code = 'bert'
        args.model_init_seed = 0

        args.model_sample_seed=0

        args.bert_hidden_units = 256
        args.bert_mask_prob = 0.15
        args.bert_max_len = seq_len
        args.bert_num_blocks = 2
        args.bert_num_heads = 4 if 'ml' in args.dataset_code else 2

        args.slide_window_step = 10 if 'ml' in args.dataset_code else 1


        args.AdaptModelHyper = AdaptModelHyper[args.dataset_code] # AdaptModel Parameter，如果每個dataset seq_len設定一樣，每個dataset的AdaptModel Parameter可以設定一樣