-
Notifications
You must be signed in to change notification settings - Fork 0
/
train.py
110 lines (100 loc) · 5 KB
/
train.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
import sys
from torch.utils.data import DataLoader
from transformers import AdamW, get_linear_schedule_with_warmup
from tqdm import tqdm
from ic_utils import *
from module import *
from dataset import StylerDALLESLDataset
from dall_e import map_pixels, unmap_pixels, load_model
def train(train_dataset, val_dataset, model, args, warmup_steps: int = 5000, exp_dir: str = "."):
device = torch.device('cuda:0')
dalle_dec = load_model('dalle/decoder.pkl', device)
batch_size = args.bs
epochs = args.epochs
lr = args.lr
work_dir = args.work_dir
output_dir = os.path.join(work_dir, exp_dir)
if not os.path.exists(output_dir):
os.makedirs(output_dir)
model = model.to(device)
optimizer = AdamW(model.parameters(), lr=lr)
train_dataloader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True, num_workers=args.workers, drop_last=True)
val_dataloader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False, num_workers=args.workers, drop_last=True)
scheduler = get_linear_schedule_with_warmup(
optimizer, num_warmup_steps=warmup_steps, num_training_steps=epochs * len(train_dataloader)
)
for epoch in range(epochs):
model.train()
running_loss = 0.0
print(f">>> Training epoch {epoch}")
sys.stdout.flush()
progress = tqdm(total=len(train_dataloader), desc=exp_dir)
for idx, (tokens_16, tokens_32, image_names) in enumerate(train_dataloader):
model.zero_grad()
bs = tokens_16.size(0)
with torch.no_grad():
dalle_encodings = dalle_dec.blocks.input(F.one_hot(tokens_16.to(device).view(bs, 16, 16), num_classes=8192).permute(0, 3, 1, 2).float())
tokens = tokens_32.to(device)
outputs = model(dalle_encodings.permute(0, 2, 3, 1).view(bs, 16*16, 128))
loss = maskedNll(outputs, tokens.contiguous().view(bs, -1))
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
progress.set_postfix({"loss": loss.item()})
progress.update()
running_loss = args.bs / (args.bs + 1) * running_loss + 1 / (args.bs + 1) * loss.data
torch.save(
model.state_dict(),
os.path.join(output_dir, f"{exp_dir}_latest.pt"),
)
with open(os.path.join(output_dir, 'logs.txt'), 'a') as f:
f.write('Train: Epoch - {}, CE - \n{}\n'.format(epoch, running_loss))
progress.close()
print(f">>> Validation epoch {epoch}")
sys.stdout.flush()
progress = tqdm(total=len(val_dataloader), desc=exp_dir)
running_loss = 0.0
for idx, (tokens_16, tokens_32, image_names) in enumerate(val_dataloader):
model.eval()
with torch.no_grad():
bs = tokens_16.size(0)
dalle_encodings = dalle_dec.blocks.input(F.one_hot(tokens_16.to(device).view(bs, 16, 16), num_classes=8192).permute(0, 3, 1, 2).float())
tokens = tokens_32.to(device)
outputs = model(dalle_encodings.permute(0, 2, 3, 1).view(bs, 256, 128))
val_loss = maskedNll(outputs, tokens.contiguous().view(bs, -1))
progress.set_postfix({"loss": val_loss.item()})
progress.update()
running_loss = args.bs / (args.bs + 1) * running_loss + 1 / (args.bs + 1) * val_loss.data
with open(os.path.join(output_dir, 'logs.txt'), 'a') as f:
f.write('Val: Epoch - {}, \n{}\n'.format(epoch, running_loss))
progress.close()
if epoch % args.save_every == 0 or epoch == epochs - 1:
torch.save(
model.state_dict(),
os.path.join(output_dir, f"{exp_dir}-{epoch:03d}.pt"),
)
return model
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--train_path', default='./prep_coco/train')
parser.add_argument('--val_path', default='./prep_coco/val')
parser.add_argument('--work_dir', default='./ic')
parser.add_argument('--exp_dir', default='sl')
parser.add_argument('--lr', type=float, default=1e-4)
parser.add_argument('--epochs', type=int, default=25)
parser.add_argument('--res', type=int, default=32)
parser.add_argument('--tok_dim', type=int, default=128)
parser.add_argument('--hidden_dim', type=int, default=512)
parser.add_argument('--num_heads', type=int, default=8)
parser.add_argument('--save_every', type=int, default=1)
parser.add_argument('--bs', type=int, default=32)
parser.add_argument('--workers', type=int, default=4)
parser.add_argument('--num_layers', type=int, default=4)
args = parser.parse_args()
model = StylerDALLEModel(args.tok_dim, args.hidden_dim, args.num_heads, args.num_layers, args.res)
device = torch.device('cuda:0')
train_dataset = StylerDALLESLDataset(args.train_path)
val_dataset = StylerDALLESLDataset(args.val_path)
sys.stdout.flush()
train(train_dataset, val_dataset, model, args, exp_dir=args.exp_dir)