forked from salesforce/ALBEF
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Grounding.py
295 lines (230 loc) · 12.4 KB
/
Grounding.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
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
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
import argparse
import os
import ruamel_yaml as yaml
import numpy as np
import random
import time
import datetime
import json
from pathlib import Path
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.backends.cudnn as cudnn
import torch.distributed as dist
from torch.utils.data import DataLoader
from models.model_retrieval import ALBEF
from models.vit import interpolate_pos_embed
from models.tokenization_bert import BertTokenizer
import utils
from dataset import create_dataset, create_sampler, create_loader
from dataset.utils import collect_result, grounding_eval
from scheduler import create_scheduler
from optim import create_optimizer
from refTools.refer_python3 import REFER
from pdb import set_trace as breakpoint
def train(model, data_loader, optimizer, tokenizer, epoch, warmup_steps, device, scheduler, config):
# train
model.train()
metric_logger = utils.MetricLogger(delimiter=" ")
metric_logger.add_meter('lr', utils.SmoothedValue(window_size=1, fmt='{value:.6f}'))
metric_logger.add_meter('loss_itm', utils.SmoothedValue(window_size=1, fmt='{value:.4f}'))
metric_logger.add_meter('loss_ita', utils.SmoothedValue(window_size=1, fmt='{value:.4f}'))
header = 'Train Epoch: [{}]'.format(epoch)
print_freq = 50
step_size = 100
warmup_iterations = warmup_steps*step_size
for i,(image, text, idx) in enumerate(metric_logger.log_every(data_loader, print_freq, header)):
image = image.to(device,non_blocking=True)
idx = idx.to(device,non_blocking=True)
text_input = tokenizer(text, padding='longest', max_length=30, return_tensors="pt").to(device)
if epoch>0 or not config['warm_up']:
alpha = config['alpha']
else:
alpha = config['alpha']*min(1,i/len(data_loader))
loss_ita, loss_itm = model(image, text_input,alpha=alpha, idx=idx)
loss = loss_ita + loss_itm
optimizer.zero_grad()
loss.backward()
optimizer.step()
metric_logger.update(loss_itm=loss_itm.item())
metric_logger.update(loss_ita=loss_ita.item())
metric_logger.update(lr=optimizer.param_groups[0]["lr"])
if epoch==0 and i%step_size==0 and i<=warmup_iterations:
scheduler.step(i//step_size)
# gather the stats from all processes
metric_logger.synchronize_between_processes()
print("Averaged stats:", metric_logger.global_avg())
return {k: "{:.3f}".format(meter.global_avg) for k, meter in metric_logger.meters.items()}
def val(model, data_loader, tokenizer, device, gradcam_mode, block_num):
# test
model.eval()
metric_logger = utils.MetricLogger(delimiter=" ")
header = 'Evaluation:'
print_freq = 50
if gradcam_mode=='itm':
model.text_encoder.base_model.base_model.encoder.layer[block_num].crossattention.self.save_attention = True
result = []
for image, text, ref_ids in metric_logger.log_every(data_loader, print_freq, header):
image = image.to(device)
text_input = tokenizer(text, padding='longest', return_tensors="pt").to(device)
if gradcam_mode=='itm':
image_embeds = model.visual_encoder(image)
image_atts = torch.ones(image_embeds.size()[:-1],dtype=torch.long).to(image.device)
output = model.text_encoder(text_input.input_ids,
attention_mask = text_input.attention_mask,
encoder_hidden_states = image_embeds,
encoder_attention_mask = image_atts,
return_dict = True,
)
vl_embeddings = output.last_hidden_state[:,0,:]
vl_output = model.itm_head(vl_embeddings)
loss = vl_output[:,1].sum()
model.zero_grad()
loss.backward()
with torch.no_grad():
mask = text_input.attention_mask.view(text_input.attention_mask.size(0),1,-1,1,1)
grads = model.text_encoder.base_model.base_model.encoder.layer[block_num].crossattention.self.get_attn_gradients().detach()
cams = model.text_encoder.base_model.base_model.encoder.layer[block_num].crossattention.self.get_attention_map().detach()
cams = cams[:, :, :, 1:].reshape(image.size(0), 12, -1, 24, 24) * mask
grads = grads[:, :, :, 1:].clamp(min=0).reshape(image.size(0), 12, -1, 24, 24) * mask
gradcam = cams * grads
gradcam = gradcam.mean(1).mean(1)
elif gradcam_mode=='itc':
image_embeds = model.visual_encoder(image, register_blk=block_num)
image_feat = F.normalize(model.vision_proj(image_embeds[:,0,:]),dim=-1)
text_output = model.text_encoder(text_input.input_ids, attention_mask = text_input.attention_mask,
return_dict = True, mode = 'text')
text_embeds = text_output.last_hidden_state
text_feat = F.normalize(model.text_proj(text_embeds[:,0,:]),dim=-1)
sim = image_feat@text_feat.t()/model.temp
loss = sim.diag().sum()
model.zero_grad()
loss.backward()
with torch.no_grad():
grad = model.visual_encoder.blocks[block_num].attn.get_attn_gradients().detach()
cam = model.visual_encoder.blocks[block_num].attn.get_attention_map().detach()
cam = cam[:, :, 0, 1:].reshape(image.size(0), -1, 24, 24)
grad = grad[:, :, 0, 1:].reshape(image.size(0), -1, 24, 24).clamp(0)
gradcam = (cam * grad).mean(1)
for r_id, cam in zip(ref_ids, gradcam):
result.append({'ref_id':r_id.item(), 'pred':cam})
if gradcam_mode=='itm':
model.text_encoder.base_model.base_model.encoder.layer[block_num].crossattention.self.save_attention = False
return result
def main(args, config):
utils.init_distributed_mode(args)
device = torch.device(args.device)
# fix the seed for reproducibility
seed = args.seed + utils.get_rank()
torch.manual_seed(seed)
np.random.seed(seed)
random.seed(seed)
cudnn.benchmark = True
#### Dataset ####
print("Creating dataset")
grd_train_dataset, grd_test_dataset = create_dataset('grounding', config)
datasets = [grd_train_dataset, grd_test_dataset]
if args.distributed:
num_tasks = utils.get_world_size()
global_rank = utils.get_rank()
samplers = create_sampler(datasets, [True, False], num_tasks, global_rank)
else:
samplers = [None, None]
train_loader, test_loader = create_loader(datasets,samplers,batch_size=[config['batch_size'],config['batch_size']],
num_workers=[4,4],is_trains=[True, False], collate_fns=[None,None])
tokenizer = BertTokenizer.from_pretrained(args.text_encoder)
## refcoco evaluation tools
refer = REFER(config['refcoco_data'], 'refcoco+', 'unc')
dets = json.load(open(config['det_file'],'r'))
cocos = json.load(open(config['coco_file'],'r'))
#### Model ####
print("Creating model")
model = ALBEF(config = config, text_encoder=args.text_encoder, tokenizer=tokenizer)
if args.checkpoint:
checkpoint = torch.load(args.checkpoint, map_location='cpu')
state_dict = checkpoint['model']
pos_embed_reshaped = interpolate_pos_embed(state_dict['visual_encoder.pos_embed'],model.visual_encoder)
state_dict['visual_encoder.pos_embed'] = pos_embed_reshaped
m_pos_embed_reshaped = interpolate_pos_embed(state_dict['visual_encoder_m.pos_embed'],model.visual_encoder_m)
state_dict['visual_encoder_m.pos_embed'] = m_pos_embed_reshaped
for key in list(state_dict.keys()):
if 'bert' in key:
encoder_key = key.replace('bert.','')
state_dict[encoder_key] = state_dict[key]
del state_dict[key]
msg = model.load_state_dict(state_dict,strict=False)
print('load checkpoint from %s'%args.checkpoint)
print(msg)
model = model.to(device)
model_without_ddp = model
if args.distributed:
model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[args.gpu])
model_without_ddp = model.module
arg_opt = utils.AttrDict(config['optimizer'])
optimizer = create_optimizer(arg_opt, model)
arg_sche = utils.AttrDict(config['schedular'])
lr_scheduler, _ = create_scheduler(arg_sche, optimizer)
max_epoch = config['schedular']['epochs']
warmup_steps = config['schedular']['warmup_epochs']
best = 0
print("Start training")
start_time = time.time()
for epoch in range(0, max_epoch):
if not args.evaluate:
if args.distributed:
train_loader.sampler.set_epoch(epoch)
train_stats = train(model, train_loader, optimizer, tokenizer, epoch, warmup_steps, device, lr_scheduler, config)
result = val(model_without_ddp, test_loader, tokenizer, device, args.gradcam_mode, args.block_num)
results = collect_result(result, args.result_dir, 'epoch%d'%epoch, is_json=False, is_list=True)
if utils.is_main_process():
grounding_acc = grounding_eval(results, dets, cocos, refer, alpha=0.5, mask_size=24)
if args.evaluate:
log_stats = {**{f'{k}': v for k, v in grounding_acc.items()},
'epoch': epoch,
}
else:
log_stats = {**{f'train_{k}': v for k, v in train_stats.items()},
**{f'{k}': v for k, v in grounding_acc.items()},
'epoch': epoch,
}
if grounding_acc['val_d']>best:
save_obj = {
'model': model_without_ddp.state_dict(),
'optimizer': optimizer.state_dict(),
'lr_scheduler': lr_scheduler.state_dict(),
'config': config,
'epoch': epoch,
}
torch.save(save_obj, os.path.join(args.output_dir, 'checkpoint_best.pth'))
best = grounding_acc['val_d']
with open(os.path.join(args.output_dir, "log.txt"),"a") as f:
f.write(json.dumps(log_stats) + "\n")
if args.evaluate:
break
lr_scheduler.step(epoch+warmup_steps+1)
dist.barrier()
total_time = time.time() - start_time
total_time_str = str(datetime.timedelta(seconds=int(total_time)))
print('Training time {}'.format(total_time_str))
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--config', default='./configs/Grounding.yaml')
parser.add_argument('--checkpoint', default='')
parser.add_argument('--output_dir', default='output/RefCOCO')
parser.add_argument('--gradcam_mode', default='itm', choices=['itm','itc'])
parser.add_argument('--block_num', default=8, type=int)
parser.add_argument('--text_encoder', default='bert-base-uncased')
parser.add_argument('--evaluate', action='store_true')
parser.add_argument('--device', default='cuda')
parser.add_argument('--seed', default=42, type=int)
parser.add_argument('--world_size', default=1, type=int, help='number of distributed processes')
parser.add_argument('--dist_url', default='env://', help='url used to set up distributed training')
parser.add_argument('--distributed', default=True, type=bool)
args = parser.parse_args()
config = yaml.load(open(args.config, 'r'), Loader=yaml.Loader)
args.result_dir = os.path.join(args.output_dir, 'result')
Path(args.output_dir).mkdir(parents=True, exist_ok=True)
Path(args.result_dir).mkdir(parents=True, exist_ok=True)
yaml.dump(config, open(os.path.join(args.output_dir, 'config.yaml'), 'w'))
main(args, config)