MoNet/train.py at main · ispc-lab/MoNet · 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
import os
import numpy as np
import torch
import torch.nn as nn
from torch.utils.data import DataLoader
import torch.optim as optim
from torch.optim.lr_scheduler import StepLR

from data.kitti_loader import KittiDataset
from data.argo_loader import ArgoDataset

from model.models import MoNet

from model.utils import batch_chamfer_distance, multi_frame_chamfer_loss, set_seed

from tqdm import tqdm
import argparse

def parse_args():
    parser = argparse.ArgumentParser(description='MoNet')

    parser.add_argument('--batch_size', type=int, default=2)
    parser.add_argument('--epochs', type=int, default=100)
    parser.add_argument('--lr', type=float, default=0.001)
    parser.add_argument('--seed', type=int, default=1)
    parser.add_argument('--gpu', type=str, default='1')
    parser.add_argument('--multi_gpu', action='store_true')
    parser.add_argument('--root', type=str, default='')
    parser.add_argument('--npoints', type=int, default=16384)
    parser.add_argument('--use_wandb', action='store_true')
    parser.add_argument('--runname', type=str, default='')
    parser.add_argument('--rnn', type=str, default='', help='LSTM/GRU')
    parser.add_argument('--pred_num', type=int, default=5)
    parser.add_argument('--input_num', type=int, default=5)
    parser.add_argument('--dataset', type=str, default='kitti')
    parser.add_argument('--ckpt_dir', type=str, default='')
    parser.add_argument('--wandb_dir', type=str, default='')

    return parser.parse_args()

def validation(args, net):
    if args.dataset == 'kitti':
        val_seqs = ['06','07']
        val_dataset = KittiDataset(args.root, args.npoints, args.input_num, args.pred_num, val_seqs)
    elif args.dataset == 'argoverse':
        val_seqs = ['val']
        val_dataset = ArgoDataset(args.root, args.npoints, args.input_num, args.pred_num, val_seqs)
    else:
        raise('Not implemented')
    val_loader = DataLoader(val_dataset,
                            batch_size=args.batch_size,
                            num_workers=4,
                            shuffle=True,
                            pin_memory=True,
                            drop_last=True)
    net.eval()

    total_val_loss = 0
    count = 0
    pbar = tqdm(enumerate(val_loader))
    with torch.no_grad():
        for i, data in pbar:
            input_pc, output_pc = data
            input_pc = input_pc.cuda()
            output_pc = output_pc.cuda()

            pred_pc = net(input_pc)

            loss = multi_frame_chamfer_loss(output_pc[:,:,:3,:], pred_pc)
            total_val_loss += loss.item()
            count += 1

    total_val_loss = total_val_loss/count
    return total_val_loss

def train(args):
    if args.dataset == 'kitti':
        train_seqs = ['00','01','02','03','04','05']
        train_dataset = KittiDataset(args.root, args.npoints, args.input_num, args.pred_num, train_seqs)
    elif args.dataset == 'argoverse':
        train_seqs = ['train1', 'train2', 'train3', 'train4']
        train_dataset = ArgoDataset(args.root, args.npoints, args.input_num, args.pred_num, train_seqs)
    else:
        raise('Not implemented')

    train_loader = DataLoader(train_dataset,
                              batch_size=args.batch_size,
                              num_workers=4,
                              shuffle=True,
                              pin_memory=True,
                              drop_last=True)

    net = MoNet(args)

    if args.use_wandb:
        wandb.watch(net)
    if args.multi_gpu:
        net = torch.nn.DataParallel(net)
    net.cuda()

    optimizer = optim.Adam(net.parameters(), lr=args.lr)
    scheduler = StepLR(optimizer, step_size=10, gamma=0.5)

    best_train_loss = float('inf')
    best_val_loss = float('inf')
    best_train_epoch = 0
    best_val_epoch = 0

    for epoch in tqdm(range(args.epochs)):

        net.train()
        count = 0
        total_loss = 0
        pbar = tqdm(enumerate(train_loader))

        for i, data in pbar:
            input_pc, output_pc = data
            input_pc = input_pc.cuda()
            output_pc = output_pc.cuda()

            optimizer.zero_grad()
            pred_pc = net(input_pc)

            loss = multi_frame_chamfer_loss(output_pc[:,:,:3,:], pred_pc)
            loss.backward()
            torch.nn.utils.clip_grad_norm_(net.parameters(),max_norm=5.0)
            optimizer.step()

            count += 1
            total_loss += loss.item()

            if i % 10 == 0:
                pbar.set_description('Train Epoch:{}[{}/{}({:.0f}%)]\tLoss: {:.6f}'.format(
                    epoch+1, i, len(train_loader), 100. * i/len(train_loader), loss.item()
                ))

        total_loss = total_loss/count
        total_val_loss = validation(args, net)

        if args.use_wandb:
            wandb.log({"train loss":total_loss, "val loss":total_val_loss})

        print('\n Epoch {} finished. Training loss: {:.4f} Valiadation loss: {:.4f}'.\
            format(epoch+1, total_loss, total_val_loss))

        ckpt_dir = os.path.join(args.ckpt_dir, 'ckpt_'+args.runname)
        if not os.path.exists(ckpt_dir):
            os.makedirs(ckpt_dir)

        if total_loss < best_train_loss:
            if args.multi_gpu:
                torch.save(net.module.state_dict(), os.path.join(ckpt_dir, 'best_train.pth'))
            else:
                torch.save(net.state_dict(), os.path.join(ckpt_dir, 'best_train.pth'))
            best_train_loss = total_loss
            best_train_epoch = epoch + 1

        if total_val_loss < best_val_loss:
            if args.multi_gpu:
                torch.save(net.module.state_dict(), os.path.join(ckpt_dir, 'best_val.pth'))
            else:
                torch.save(net.state_dict(), os.path.join(ckpt_dir, 'best_val.pth'))
            best_val_loss = total_val_loss
            best_val_epoch = epoch + 1

        print('Best train epoch: {} Best train loss: {:.4f} Best val epoch: {} Best val loss: {:.4f}'.format(
            best_train_epoch, best_train_loss, best_val_epoch, best_val_loss
        ))
        scheduler.step()

if __name__ == '__main__':
    args = parse_args()

    os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
    set_seed(args.seed)

    if args.use_wandb:
        import wandb
        wandb.init(config=args, project='MoNet', name=args.dataset+'_'+args.runname, dir=args.wandb_dir)
    train(args)