A-LLMRec/train_model.py at master · mshadloo/A-LLMRec · 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
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
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
import os
import torch
import random
import time
import os

from tqdm import tqdm

import torch.multiprocessing as mp
from torch.utils.data import DataLoader
from torch.utils.data.distributed import DistributedSampler
from torch.nn.parallel import DistributedDataParallel as DDP
from torch.distributed import init_process_group, destroy_process_group

from models.a_llmrec_model import *
from pre_train.sasrec.utils import data_partition, SeqDataset, SeqDataset_Inference


def setup_ddp(rank, world_size):
    os.environ ["MASTER_ADDR"] = "localhost"
    os.environ ["MASTER_PORT"] = "12355"
    init_process_group(backend="nccl", rank=rank, world_size=world_size)
    torch.cuda.set_device(rank)

def train_model_phase1(args):
    print('A-LLMRec start train phase-1\n')
    if args.multi_gpu:
        world_size = torch.cuda.device_count()
        mp.spawn(train_model_phase1_, args=(world_size, args), nprocs=world_size)
    else:
        train_model_phase1_(0, 0, args)

def train_model_phase2(args):
    print('A-LLMRec strat train phase-2\n')
    if args.multi_gpu:
        world_size = torch.cuda.device_count()
        mp.spawn(train_model_phase2_, args=(world_size, args), nprocs=world_size)
    else:
        train_model_phase2_(0, 0, args)

def inference(args):
    print('A-LLMRec start inference\n')
    if args.multi_gpu:
        world_size = torch.cuda.device_count()
        mp.spawn(inference_, args=(world_size, args), nprocs=world_size)
    else:
        inference_(0,0,args)

def train_model_phase1_(rank, world_size, args):
    if args.multi_gpu:
        setup_ddp(rank, world_size)
        args.device = 'cuda:' + str(rank)

    model = A_llmrec_model(args).to(args.device)

    # preprocess data
    dataset = data_partition(args.rec_pre_trained_data, path=f'./data/amazon/{args.rec_pre_trained_data}.txt')
    [user_train, user_valid, user_test, usernum, itemnum] = dataset
    print('user num:', usernum, 'item num:', itemnum)
    num_batch = len(user_train) // args.batch_size1
    cc = 0.0
    for u in user_train:
        cc += len(user_train[u])
    print('average sequence length: %.2f' % (cc / len(user_train)))
    # Init Dataloader, Model, Optimizer
    train_data_set = SeqDataset(user_train, usernum, itemnum, args.maxlen)
    if args.multi_gpu:
        train_data_loader = DataLoader(train_data_set, batch_size = args.batch_size1, sampler=DistributedSampler(train_data_set, shuffle=True), pin_memory=True)
        model = DDP(model, device_ids = [args.device], static_graph=True)
    else:
        train_data_loader = DataLoader(train_data_set, batch_size = args.batch_size1, pin_memory=True)

    adam_optimizer = torch.optim.Adam(model.parameters(), lr=args.stage1_lr, betas=(0.9, 0.98))

    epoch_start_idx = 1
    T = 0.0
    model.train()
    t0 = time.time()
    for epoch in tqdm(range(epoch_start_idx, args.num_epochs + 1)):
        if args.multi_gpu:
            train_data_loader.sampler.set_epoch(epoch)
        for step, data in enumerate(train_data_loader):
            u, seq, pos, neg = data
            u, seq, pos, neg = u.numpy(), seq.numpy(), pos.numpy(), neg.numpy()
            model([u,seq,pos,neg], optimizer=adam_optimizer, batch_iter=[epoch,args.num_epochs + 1,step,num_batch], mode='phase1')
            if step % max(10,num_batch//100) ==0:
                if rank ==0:
                    if args.multi_gpu: model.module.save_model(args, epoch1=epoch)
                    else: model.save_model(args, epoch1=epoch)
        if rank == 0:
            if args.multi_gpu: model.module.save_model(args, epoch1=epoch)
            else: model.save_model(args, epoch1=epoch)

    print('train time :', time.time() - t0)
    if args.multi_gpu:
        destroy_process_group()
    return

def train_model_phase2_(rank,world_size,args):
    if args.multi_gpu:
        setup_ddp(rank, world_size)
        args.device = 'cuda:'+str(rank)
    random.seed(0)

    model = A_llmrec_model(args).to(args.device)
    phase1_epoch = 10
    model.load_model(args, phase1_epoch=phase1_epoch)

    dataset = data_partition(args.rec_pre_trained_data, path=f'./data/amazon/{args.rec_pre_trained_data}.txt')
    [user_train, user_valid, user_test, usernum, itemnum] = dataset
    print('user num:', usernum, 'item num:', itemnum)
    num_batch = len(user_train) // args.batch_size2
    cc = 0.0
    for u in user_train:
        cc += len(user_train[u])
    print('average sequence length: %.2f' % (cc / len(user_train)))
    # Init Dataloader, Model, Optimizer
    train_data_set = SeqDataset(user_train, usernum, itemnum, args.maxlen)
    if args.multi_gpu:
        train_data_loader = DataLoader(train_data_set, batch_size = args.batch_size2, sampler=DistributedSampler(train_data_set, shuffle=True), pin_memory=True)
        model = DDP(model, device_ids = [args.device], static_graph=True)
    else:
        train_data_loader = DataLoader(train_data_set, batch_size = args.batch_size2, pin_memory=True, shuffle=True)
    adam_optimizer = torch.optim.Adam(model.parameters(), lr=args.stage2_lr, betas=(0.9, 0.98))

    epoch_start_idx = 1
    T = 0.0
    model.train()
    t0 = time.time()
    for epoch in tqdm(range(epoch_start_idx, args.num_epochs + 1)):
        if args.multi_gpu:
            train_data_loader.sampler.set_epoch(epoch)
        for step, data in enumerate(train_data_loader):
            u, seq, pos, neg = data
            u, seq, pos, neg = u.numpy(), seq.numpy(), pos.numpy(), neg.numpy()
            model([u,seq,pos,neg], optimizer=adam_optimizer, batch_iter=[epoch,args.num_epochs + 1,step,num_batch], mode='phase2')
            if step % max(10,num_batch//100) ==0:
                if rank ==0:
                    if args.multi_gpu: model.module.save_model(args, epoch1=phase1_epoch, epoch2=epoch)
                    else: model.save_model(args, epoch1=phase1_epoch, epoch2=epoch)
        if rank == 0:
            if args.multi_gpu: model.module.save_model(args, epoch1=phase1_epoch, epoch2=epoch)
            else: model.save_model(args, epoch1=phase1_epoch, epoch2=epoch)

    print('phase2 train time :', time.time() - t0)
    if args.multi_gpu:
        destroy_process_group()
    return

def inference_(rank, world_size, args):
    if args.multi_gpu:
        setup_ddp(rank, world_size)
        args.device = 'cuda:' + str(rank)

    model = A_llmrec_model(args).to(args.device)
    phase1_epoch = 10
    phase2_epoch = 5
    model.load_model(args, phase1_epoch=phase1_epoch, phase2_epoch=phase2_epoch)

    dataset = data_partition(args.rec_pre_trained_data, path=f'./data/amazon/{args.rec_pre_trained_data}.txt')
    [user_train, user_valid, user_test, usernum, itemnum] = dataset
    print('user num:', usernum, 'item num:', itemnum)
    num_batch = len(user_train) // args.batch_size_infer
    cc = 0.0
    for u in user_train:
        cc += len(user_train[u])
    print('average sequence length: %.2f' % (cc / len(user_train)))
    model.eval()

    if usernum>10000:
        users = random.sample(range(1, usernum + 1), 10000)
    else:
        users = range(1, usernum + 1)

    user_list = []
    for u in users:
        if len(user_train[u]) < 1 or len(user_test[u]) < 1: continue
        user_list.append(u)

    inference_data_set = SeqDataset_Inference(user_train, user_valid, user_test, user_list, itemnum, args.maxlen)

    if args.multi_gpu:
        inference_data_loader = DataLoader(inference_data_set, batch_size = args.batch_size_infer, sampler=DistributedSampler(inference_data_set, shuffle=True), pin_memory=True)
        model = DDP(model, device_ids = [args.device], static_graph=True)
    else:
        inference_data_loader = DataLoader(inference_data_set, batch_size = args.batch_size_infer, pin_memory=True)

    for _, data in enumerate(inference_data_loader):
        u, seq, pos, neg = data
        u, seq, pos, neg = u.numpy(), seq.numpy(), pos.numpy(), neg.numpy()
        model([u,seq,pos,neg, rank], mode='generate')