-
Notifications
You must be signed in to change notification settings - Fork 24
/
run_transductive_cnaps_mt.py
246 lines (208 loc) · 13.2 KB
/
run_transductive_cnaps_mt.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
import torch
import numpy as np
import argparse
import os
import sys
from utils import print_and_log, get_log_files, ValidationAccuracies, loss, aggregate_accuracy
from transductive_cnaps import TransductiveCnaps
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' # Quiet TensorFlow warnings
import tensorflow as tf
tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR) # Quiet TensorFlow warnings
from miniimagenettools.mini_imagenet_dataloader import MiniImageNetDataLoader
from tieredimagenettools.tiered_imagenet_dataloader import TieredImageNetDataLoader
NUM_TRAIN_TASKS = 20000
NUM_VALIDATION_TASKS = 200
NUM_TEST_TASKS = 600
VALIDATION_FREQUENCY = 1000
def main():
learner = Learner()
learner.run()
class Learner:
def __init__(self):
self.args = self.parse_command_line()
self.checkpoint_dir, self.logfile, self.checkpoint_path_validation, self.checkpoint_path_final \
= get_log_files(self.args.checkpoint_dir)
print_and_log(self.logfile, "Options: %s\n" % self.args)
print_and_log(self.logfile, "Checkpoint Directory: %s\n" % self.checkpoint_dir)
gpu_device = 'cuda:0'
self.device = torch.device(gpu_device if torch.cuda.is_available() else 'cpu')
self.model = self.init_model()
if self.args.dataset == "mini":
self.imagenet_subset = MiniImageNetDataLoader(shot_num=self.args.shots, way_num=self.args.ways, episode_test_sample_num=self.args.testnum)
else:
self.imagenet_subset = TieredImageNetDataLoader(shot_num=self.args.shots, way_num=self.args.ways, episode_test_sample_num=self.args.testnum)
self.imagenet_subset.generate_data_list(phase='train')
self.imagenet_subset.generate_data_list(phase='val')
self.imagenet_subset.generate_data_list(phase='test')
self.imagenet_subset.load_list(phase='all')
self.loss = loss
self.accuracy_fn = aggregate_accuracy
self.optimizer = torch.optim.Adam(self.model.parameters(), lr=self.args.learning_rate)
self.optimizer.zero_grad()
def init_model(self):
use_two_gpus = self.use_two_gpus()
model = TransductiveCnaps(device=self.device, use_two_gpus=use_two_gpus, args=self.args, mt=True).to(self.device)
model.train() # set encoder is always in train mode to process context data
model.feature_extractor.eval() # feature extractor is always in eval mode
if use_two_gpus:
model.distribute_model()
return model
"""
Command line parser
"""
def parse_command_line(self):
parser = argparse.ArgumentParser()
parser.add_argument("--pretrained_resnet_path", default="../model-checkpoints/pretrained_resnets/pretrained_resnet_mini_tiered_with_extra_classes.pt.tar",
help="Path to pretrained feature extractor model.")
parser.add_argument("--mode", choices=["train", "test", "train_test"], default="train_test",
help="Whether to run training only, testing only, or both training and testing.")
parser.add_argument("--learning_rate", "-lr", type=float, default=5e-4, help="Learning rate.")
parser.add_argument("--tasks_per_batch", type=int, default=16,
help="Number of tasks between parameter optimizations.")
parser.add_argument("--checkpoint_dir", "-c", default='../checkpoints', help="Directory to save checkpoint to.")
parser.add_argument("--test_model_path", "-m", default=None, help="Path to model to load and test.")
parser.add_argument("--feature_adaptation", choices=["no_adaptation", "film", "film+ar"], default="film",
help="Method to adapt feature extractor parameters.")
parser.add_argument("--shots", type=int, default=1,
help="Number of shots in the task.")
parser.add_argument("--ways", type=int, default=5,
help="Number of ways in the task.")
parser.add_argument("--testnum", type=int, default=10,
help="Number of test examples per class in the task.")
parser.add_argument("--dataset", choices=["mini", "tiered"], default="mini",
help="Imagenet subset dataset to be used.")
# Transductive CNAPS specific parameters
parser.add_argument("--min_cluster_refinement_steps_train", type=int, default=0, help="Minimum number of cluster refinement steps to take whilst training.")
parser.add_argument("--max_cluster_refinement_steps_train", type=int, default=0, help="Maximum number of cluster refinement steps to take whilst training.")
parser.add_argument("--min_cluster_refinement_steps_test", type=int, default=2, help="Minimum number of cluster refinement steps to take.")
parser.add_argument("--max_cluster_refinement_steps_test", type=int, default=4, help="Maximum number of cluster refinement steps to take.")
args = parser.parse_args()
return args
def run(self):
config = tf.compat.v1.ConfigProto()
config.gpu_options.allow_growth = True
validation_accuracy = 0
with tf.compat.v1.Session(config=config) as session:
if self.args.mode == 'train' or self.args.mode == 'train_test':
self.model.set_cluster_refinement_param_mode_to_train()
train_accuracies = []
losses = []
total_iterations = NUM_TRAIN_TASKS
for iteration in range(total_iterations):
torch.set_grad_enabled(True)
task_loss, task_accuracy = self.train_task(iteration)
train_accuracies.append(task_accuracy)
losses.append(task_loss)
# optimize
if ((iteration + 1) % self.args.tasks_per_batch == 0) or (iteration == (total_iterations - 1)):
self.optimizer.step()
self.optimizer.zero_grad()
if (iteration + 1) % 250 == 0:
# print training stats
print_and_log(self.logfile,'Task [{}/{}], Train Loss: {:.7f}, Train Accuracy: {:.7f}'
.format(iteration + 1, total_iterations, torch.Tensor(losses).mean().item(),
torch.Tensor(train_accuracies).mean().item()))
train_accuracies = []
losses = []
if ((iteration + 1) % VALIDATION_FREQUENCY == 0) and (iteration + 1) != total_iterations:
# validate
accuracy = self.validate(session)
# save the model if validation is the best so far
if accuracy > validation_accuracy:
validation_accuracy = accuracy
torch.save(self.model.state_dict(), self.checkpoint_path_validation)
print('Best validation model was updated.')
# save the final model
torch.save(self.model.state_dict(), self.checkpoint_path_final)
if self.args.mode == 'train_test':
print("Evaluating with train mode max/min refinement steps.")
print("Max refinement steps (train): " + str(self.args.max_cluster_refinement_steps_train))
print("Min refinement steps (train): " + str(self.args.min_cluster_refinement_steps_train))
self.test(self.checkpoint_path_final, session, "train")
self.test(self.checkpoint_path_validation, session, "train")
print("Evaluating with test mode max/min refinement steps.")
print("Max refinement steps (test): " + str(self.args.max_cluster_refinement_step_test))
print("Min refinement steps (test): " + str(self.args.min_cluster_refinement_steps_test))
self.test(self.checkpoint_path_final, session, "test")
self.test(self.checkpoint_path_validation, session, "test")
if self.args.mode == 'test':
print("Evaluating with train mode max/min refinement steps.")
print("Max refinement steps (train): " + str(self.args.max_cluster_refinement_steps_train))
print("Min refinement steps (train): " + str(self.args.min_cluster_refinement_steps_train))
self.test(self.args.test_model_path, session, "train")
print("Evaluating with test mode max/min refinement steps.")
print("Max refinement steps (test): " + str(self.args.max_cluster_refinement_step_test))
print("Min refinement steps (test): " + str(self.args.min_cluster_refinement_steps_test))
self.test(self.args.test_model_path, session, "test")
self.logfile.close()
def train_task(self, idx, task_dict=None):
context_images, context_labels, target_images, target_labels = \
self.imagenet_subset.get_batch(phase='train', idx=idx)
context_images = torch.from_numpy(context_images).permute(0,3,1,2).cuda(0).float()
context_labels = torch.argmax(torch.from_numpy(context_labels), dim=1).cuda(0)
target_images = torch.from_numpy(target_images).permute(0,3,1,2).cuda(0).float()
target_labels = torch.argmax(torch.from_numpy(target_labels), dim=1).cuda(0)
target_logits = self.model(context_images, context_labels, target_images)
task_loss = self.loss(target_logits, target_labels, self.device) / self.args.tasks_per_batch
if self.args.feature_adaptation == 'film' or self.args.feature_adaptation == 'film+ar':
if self.use_two_gpus():
regularization_term = (self.model.feature_adaptation_network.regularization_term()).cuda(0)
else:
regularization_term = (self.model.feature_adaptation_network.regularization_term())
regularizer_scaling = 0.001
task_loss += regularizer_scaling * regularization_term
task_accuracy = self.accuracy_fn(target_logits, target_labels)
task_loss.backward(retain_graph=False)
return task_loss, task_accuracy
def validate(self, session):
with torch.no_grad():
accuracies = []
for idx in range(NUM_VALIDATION_TASKS):
context_images, context_labels, target_images, target_labels = \
self.imagenet_subset.get_batch(phase='val', idx=idx)
context_images = torch.from_numpy(context_images).permute(0,3,1,2).cuda(0).float()
context_labels = torch.argmax(torch.from_numpy(context_labels), dim=1).cuda(0)
target_images = torch.from_numpy(target_images).permute(0,3,1,2).cuda(0).float()
target_labels = torch.argmax(torch.from_numpy(target_labels), dim=1).cuda(0)
target_logits = self.model(context_images, context_labels, target_images)
accuracy = self.accuracy_fn(target_logits, target_labels)
accuracies.append(accuracy.item())
del target_logits
accuracy = np.array(accuracies).mean() * 100.0
confidence = (196.0 * np.array(accuracies).std()) / np.sqrt(len(accuracies))
print("Validation Accuracy:", accuracy, "Confidence:", confidence)
return accuracy
def test(self, path, session, cluster_refinement_param_mode):
self.model = self.init_model()
self.model.load_state_dict(torch.load(path))
if cluster_refinement_param_mode == "train":
self.model.set_cluster_refinement_param_mode_to_train()
elif cluster_refinement_param_mode == "test":
self.model.set_cluster_refinement_param_mode_to_test()
else:
raise Exception("Model mode should be one of 'train' or 'test'.")
print_and_log(self.logfile, "") # add a blank line
print_and_log(self.logfile, 'Testing model {0:}: '.format(path))
with torch.no_grad():
accuracies = []
for idx in range(NUM_TEST_TASKS):
context_images, context_labels, target_images, target_labels = \
self.imagenet_subset.get_batch(phase='test', idx=idx)
context_images = torch.from_numpy(context_images).permute(0,3,1,2).cuda(0).float()
context_labels = torch.argmax(torch.from_numpy(context_labels), dim=1).cuda(0)
target_images = torch.from_numpy(target_images).permute(0,3,1,2).cuda(0).float()
target_labels = torch.argmax(torch.from_numpy(target_labels), dim=1).cuda(0)
target_logits = self.model(context_images, context_labels, target_images)
accuracy = self.accuracy_fn(target_logits, target_labels)
accuracies.append(accuracy.item())
del target_logits
accuracy = np.array(accuracies).mean() * 100.0
confidence = (196.0 * np.array(accuracies).std()) / np.sqrt(len(accuracies))
print("Test Accuracy:", accuracy, "Confidence:", confidence)
def use_two_gpus(self):
use_two_gpus = False
if self.args.feature_adaptation == "film+ar":
use_two_gpus = True # film+ar model does not fit on one GPU, so use model parallelism
return use_two_gpus
if __name__ == "__main__":
main()