-
Notifications
You must be signed in to change notification settings - Fork 8
/
Copy pathtest.py
188 lines (155 loc) · 7.75 KB
/
test.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
import os
import clip
import torch.nn as nn
from datasets import Breakfast, GTEA, SALADS
from torch.utils.data import DataLoader
from tqdm import tqdm
import wandb
import argparse
import shutil
from pathlib import Path
import yaml
from dotmap import DotMap
import pprint
import numpy
from modules.fusion_module import fusion_earlyhyp
from utils.Augmentation import get_augmentation
import torch
from utils.text_prompt import *
class TextCLIP(nn.Module):
def __init__(self, model):
super(TextCLIP, self).__init__()
self.model = model
def forward(self, text):
return self.model.encode_text(text)
class ImageCLIP(nn.Module):
def __init__(self, model):
super(ImageCLIP, self).__init__()
self.model = model
def forward(self, image):
return self.model.encode_image(image)
def validate(epoch, val_loader, device, model, fusion_model, config,
text_dict_cnts, text_dict_acts, text_dict_posemb, num_aug, cnt_max):
model.eval()
fusion_model.eval()
final_act_1 = []
final_act_5 = []
final_cnt = []
gt_act = []
with torch.no_grad():
text_inputs_cnts = text_dict_cnts.to(device)
text_dict_posemb = text_dict_posemb.to(device)
text_dict_acts = text_dict_acts.to(device)
text_features_cnts = model.encode_text(text_inputs_cnts)
text_dict_posemb = model.encode_text(text_dict_posemb)
text_features_acts = model.encode_text(text_dict_acts).view(cnt_max, -1, text_dict_posemb.shape[-1])
for iii, (image, class_id) in enumerate(tqdm(val_loader)):
image = image.view((-1, config.data.num_frames, 3) + image.size()[-2:])
b, t, c, h, w = image.size()
text_features_posemb = text_dict_posemb.repeat(b, 1, 1)
gt_act.append(class_id.numpy())
image_input = image.to(device).view(-1, c, h, w)
image_features = model.encode_image(image_input).view(b, t, -1)
image_features = image_features.unsqueeze(1).repeat(1, cnt_max, 1, 1)
cnt_emb, image_features = fusion_model(image_features, text_features_posemb)
image_features /= image_features.norm(dim=-1, keepdim=True)
cnt_emb /= cnt_emb.norm(dim=-1, keepdim=True)
text_features_cnts /= text_features_cnts.norm(dim=-1, keepdim=True)
text_features_acts /= text_features_acts.norm(dim=-1, keepdim=True)
similarity_cnts = (100.0 * cnt_emb @ text_features_cnts.T)
similarity_cnts = similarity_cnts.view(b, -1).softmax(dim=-1)
_, indices_cnts = similarity_cnts.topk(1, dim=-1)
final_ind_1 = torch.zeros(b, cnt_max).long().to(device)
final_ind_5 = torch.zeros(b, cnt_max, 5).long().to(device)
for i in range(text_features_acts.shape[0]):
similarity_acts = (100.0 * image_features[:, i, :] @ text_features_acts[i, :].T)
similarity_acts = similarity_acts.view(b, num_aug, -1).softmax(dim=-1)
similarity_acts = similarity_acts.mean(dim=1, keepdim=False)
values_1, indices_1 = similarity_acts.topk(1, dim=-1)
values_5, indices_5 = similarity_acts.topk(5, dim=-1)
indices_1 = torch.where(indices_1 == 19, -1, indices_1)
indices_5 = torch.where(indices_5 == 19, -1, indices_5)
final_ind_1[:, i] = indices_1.squeeze()
final_ind_5[:, i, :] = indices_5
final_act_1.append(final_ind_1.cpu().numpy())
final_act_5.append(final_ind_5.cpu().numpy())
final_cnt.append(indices_cnts.cpu().numpy())
np.save(f'./prompt_test/gtea/split{config.data.n_split}/final_act_1.npy',
np.vstack(final_act_1))
np.save(f'./prompt_test/gtea/split{config.data.n_split}/final_act_5.npy',
np.vstack(final_act_5))
np.save(f'./prompt_test/gtea/split{config.data.n_split}/final_cnt_1.npy',
np.vstack(final_cnt))
np.save(f'./prompt_test/gtea/split{config.data.n_split}/gt_act.npy',
np.vstack(gt_act))
def main():
global args, best_prec1
global global_step
parser = argparse.ArgumentParser()
parser.add_argument('--config', '-cfg', default='./configs/gtea/gtea_test.yaml')
parser.add_argument('--log_time', default='')
parser.add_argument('--dataset', default='gtea')
args = parser.parse_args()
with open(args.config, 'r') as f:
config = yaml.load(f)
working_dir = os.path.join('./exp', config['network']['type'], config['network']['arch'], config['data']['dataset'],
args.log_time)
print('-' * 80)
print(' ' * 20, "working dir: {}".format(working_dir))
print('-' * 80)
print('-' * 80)
print(' ' * 30, "Config")
pp = pprint.PrettyPrinter(indent=4)
pp.pprint(config)
print('-' * 80)
config = DotMap(config)
Path(working_dir).mkdir(parents=True, exist_ok=True)
shutil.copy(args.config, working_dir)
shutil.copy('test.py', working_dir)
device = "cuda" if torch.cuda.is_available() else "cpu" # If using GPU then use mixed precision training.
base_model, model_state_dict = clip.load(config.network.arch, device=device, jit=False, tsm=config.network.tsm,
T=config.data.num_segments, dropout=config.network.drop_out,
emb_dropout=config.network.emb_dropout) # Must set jit=False for training ViT-B/32
transform_val = get_augmentation(False, config)
fusion_model = fusion_earlyhyp(config.network.sim_header, model_state_dict, config.data.num_frames)
model_text = TextCLIP(base_model)
model_image = ImageCLIP(base_model)
model_text = torch.nn.DataParallel(model_text).cuda()
model_image = torch.nn.DataParallel(model_image).cuda()
fusion_model = torch.nn.DataParallel(fusion_model).cuda()
if args.dataset == 'breakfast':
val_data = Breakfast(transform=transform_val, mode='val', num_frames=config.data.num_frames,
ds=config.data.ds, ol=config.data.ol)
elif args.dataset == 'gtea':
val_data = GTEA(transform=transform_val, mode='val', num_frames=config.data.num_frames,
n_split=config.data.n_split)
elif args.dataset == 'salads':
val_data = SALADS(transform=transform_val, mode='val', num_frames=config.data.num_frames,
n_split=config.data.n_split)
val_loader = DataLoader(val_data, batch_size=config.data.batch_size, num_workers=config.data.workers, shuffle=False,
pin_memory=True, drop_last=False)
if device == "cpu":
model_text.float()
model_image.float()
else:
clip.model.convert_weights(
model_text) # Actually this line is unnecessary since clip by default already on float16
clip.model.convert_weights(model_image)
start_epoch = config.solver.start_epoch
if config.pretrain:
if os.path.isfile(config.pretrain):
print(("=> loading checkpoint '{}'".format(config.pretrain)))
checkpoint = torch.load(config.pretrain)
base_model.load_state_dict(checkpoint['model_state_dict'])
fusion_model.load_state_dict(checkpoint['fusion_model_state_dict'])
del checkpoint
else:
print(("=> no checkpoint found at '{}'".format(config.pretrain)))
cnt_max = 6
text_dict_posemb = text_prompt_ord_emb(cnt_max=cnt_max)
text_dict_cnts, text_dict_acts, num_aug = text_prompt_slide_val_all(val_data.classes, cnt_max=cnt_max)
best_prec1 = 0.0
validate(start_epoch, val_loader, device, base_model, fusion_model, config,
text_dict_cnts, text_dict_acts, text_dict_posemb, num_aug, cnt_max)
if __name__ == '__main__':
main()