File size: 6,442 Bytes
2d13eb2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
import argparse
import numpy as np
import pandas as pd
import math
import random
import os
from scipy.sparse import csr_matrix
import torch
import ipdb


def get_args():
    parser = argparse.ArgumentParser()

    parser.add_argument('--data_dir', default='../../data/', type=str)
    parser.add_argument('--output_dir', default='./models/', type=str)
    parser.add_argument('--data_name', default='games', type=str)
    parser.add_argument("--gpu_id", type=str, default="0", help="gpu_id")
    parser.add_argument("--no_cuda", action="store_true")
    parser.add_argument("--seed", default=42, type=int)
    parser.add_argument('--log_name',type=str, default='log')

    # surrogate model args (SASRec) 
    parser.add_argument("--model_name", default='SASRec', type=str)
    parser.add_argument("--hidden_size", type=int, default=64,help="hidden size of transformer model")
    parser.add_argument("--num_hidden_layers", type=int, default=2, help="number of layers")
    parser.add_argument('--num_attention_heads', default=2, type=int) # 1,2,3
    parser.add_argument('--hidden_act', default="gelu", type=str)  # gelu relu
    parser.add_argument("--attention_probs_dropout_prob", type=float, default=0.5,help="attention dropout p")
    parser.add_argument("--hidden_dropout_prob", type=float, default=0.5,help="hidden dropout p")
    parser.add_argument("--initializer_range", type=float, default=0.02)
    parser.add_argument('--max_seq_length', default=50, type=int)

    # train args
    parser.add_argument('--do_eval', action='store_true', help='testing mode') 
    parser.add_argument("--batch_size", type=int, default=500, help="number of batch_size")
    parser.add_argument("--epochs", type=int, default=200, help="number of epochs")
    parser.add_argument("--topN", default='[5, 10, 20, 50]', help="the recommended item num")
    
    # optimization args
    parser.add_argument("--lr", type=float, default=0.001, help="learning rate of adam")
    parser.add_argument("--weight_decay", type=float, default=0.0, help="weight_decay of adam")
    parser.add_argument("--adam_beta1", type=float, default=0.9, help="adam first beta value")
    parser.add_argument("--adam_beta2", type=float, default=0.999, help="adam second beta value")
    
    # LLM args
    parser.add_argument("--base_model", type=str, default="/storage_fast/rhshui/llm/llama_hf/7B/", help="your path to hf llama model weight")
    parser.add_argument("--train_data_path", type=str, default="/storage/xylin/LLM4Rec/efficient_tuning/code/DEALRec/data/games/train_49156.json", help="your path of the training data")
    parser.add_argument("--resume_from_checkpoint", type=str, default="/storage/xylin/recommendation/LLM/code/alpaca/alpaca-lora-7B/", help="path of the alpaca lora adapter")
    parser.add_argument("--cutoff_len", default=512, type=int, help="cut off length for LLM input")
    parser.add_argument("--lora_r", default=8, type=int, help="lora r")
    parser.add_argument("--lora_alpha", default=16, help="lora alpha")
    parser.add_argument("--lora_dropout", default=0.05, help="lora dropout")

    # DEALRec args
    parser.add_argument('--n_fewshot', default=1024, type=int)
    parser.add_argument("--lamda", type=float, default=0.5, help="strength of gap regularization (effort score)")
    parser.add_argument('--k', default=25, type=int, help="number of groups")
    parser.add_argument("--hard_prune", default=0.1, type=float, help='percentage of hard samples to prune at first')
    parser.add_argument("--iteration", default=1, type=int, help="number of iteration to get the averaged HVP estimation")
    parser.add_argument("--recursion_depth", type=int, default=5000, help="number of recursion depth to get the h_estimation")
    args = parser.parse_args()
    return args

def set_seed(seed):
    random.seed(seed)
    os.environ['PYTHONHASHSEED'] = str(seed)
    np.random.seed(seed)
    torch.manual_seed(seed)
    torch.cuda.manual_seed(seed)
    torch.cuda.manual_seed_all(seed)
    # some cudnn methods can be random even after fixing the seed
    # unless you tell it to be deterministic
    torch.backends.cudnn.deterministic = True


def check_path(path):
    if not os.path.exists(path):
        os.makedirs(path)
        print(f'{path} created')


def neg_sample(item_set, item_size):  # 前闭后闭
    item = random.randint(1, item_size - 1)
    while item in item_set:
        item = random.randint(1, item_size - 1)
    return item

def generate_rating_matrix_valid(user_seq, num_users, num_items):
    # three lists are used to construct sparse matrix
    row = []
    col = []
    data = []
    for user_id, item_list in enumerate(user_seq):
        for item in item_list:  #
            row.append(user_id)
            col.append(item)
            data.append(1)

    row = np.array(row)
    col = np.array(col)
    data = np.array(data)
    rating_matrix = csr_matrix((data, (row, col)), shape=(num_users, num_items))

    return rating_matrix

def generate_rating_matrix_test(user_seq, user_seq_tr, num_users, num_items):
    # three lists are used to construct sparse matrix
    row = []
    col = []
    data = []
    for user_id, item_list in enumerate(user_seq):
        for item in item_list:  #
            row.append(user_id)
            col.append(item)
            data.append(1)
    for user_id, item_list in enumerate(user_seq_tr):
        for item in item_list:
            row.append(user_id)
            col.append(item)
            data.append(1)

    row = np.array(row)
    col = np.array(col)
    data = np.array(data)
    rating_matrix = csr_matrix((data, (row, col)), shape=(num_users, num_items))

    return rating_matrix

def get_user_seqs_npy(data_file):
    print(data_file)
    data_dict = np.load(data_file,allow_pickle=True).item()
    data_dict = dict(sorted(data_dict.items()))
    user_seq = []
    max_item = 0
    for u_id in data_dict:
        for i,i_id in enumerate(data_dict[u_id]):
            data_dict[u_id][i] = i_id + 1
        user_seq.append(data_dict[u_id])
        if len(data_dict[u_id])!=0:
            max_item = max(max_item, max(data_dict[u_id]))
    return user_seq 

def get_statistics(path):
    item2id = np.load(path + "/item_map.npy", allow_pickle=True).item()
    user2id = np.load(path + "/user_map.npy", allow_pickle=True).item()
    max_item, num_users, num_items = len(item2id), len(user2id), len(item2id)+1
    return max_item, num_users, num_items