-
Notifications
You must be signed in to change notification settings - Fork 13
/
train.py
464 lines (389 loc) · 20.3 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
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
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
#This code is taken from Mika's PointnetVLAD
import argparse
import math
import numpy as np
import tensorflow as tf
import socket
import importlib
import os
import sys
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
sys.path.append(BASE_DIR)
from pcan_cls import *
from loading_pointclouds import *
from sklearn.neighbors import NearestNeighbors
from sklearn.neighbors import KDTree
#params
parser = argparse.ArgumentParser()
parser.add_argument('--gpu', type=int, default=0, help='GPU to use [default: GPU 1]')
parser.add_argument('--log_dir', default='log/', help='Log dir [default: log]')
parser.add_argument('--positives_per_query', type=int, default=2, help='Number of potential positives in each training tuple [default: 2]')
parser.add_argument('--negatives_per_query', type=int, default=8, help='Number of definite negatives in each training tuple [default: 18]')
parser.add_argument('--max_epoch', type=int, default=20, help='Epoch to run [default: 20]')
parser.add_argument('--batch_num_queries', type=int, default=1, help='Batch Size during training [default: 2]')
parser.add_argument('--learning_rate', type=float, default=0.00005, help='Initial learning rate [default: 0.00005]')
parser.add_argument('--momentum', type=float, default=0.9, help='Initial learning rate [default: 0.9]')
parser.add_argument('--optimizer', default='adam', help='adam or momentum [default: adam]')
parser.add_argument('--decay_step', type=int, default=200000, help='Decay step for lr decay [default: 200000]')
parser.add_argument('--decay_rate', type=float, default=0.7, help='Decay rate for lr decay [default: 0.7]')
parser.add_argument('--margin_1', type=float, default=0.5, help='Margin for hinge loss [default: 0.5]')
parser.add_argument('--margin_2', type=float, default=0.2, help='Margin for hinge loss [default: 0.2]')
FLAGS = parser.parse_args()
BATCH_NUM_QUERIES = FLAGS.batch_num_queries
EVAL_BATCH_SIZE = 1
NUM_POINTS = 4096
POSITIVES_PER_QUERY= FLAGS.positives_per_query
NEGATIVES_PER_QUERY= FLAGS.negatives_per_query
MAX_EPOCH = FLAGS.max_epoch
BASE_LEARNING_RATE = FLAGS.learning_rate
GPU_INDEX = FLAGS.gpu
MOMENTUM = FLAGS.momentum
OPTIMIZER = FLAGS.optimizer
DECAY_STEP = FLAGS.decay_step
DECAY_RATE = FLAGS.decay_rate
MARGIN1 = FLAGS.margin_1
MARGIN2 = FLAGS.margin_2
TRAIN_FILE = 'generating_queries/training_queries_baseline.pickle'
TEST_FILE = 'generating_queries/test_queries_baseline.pickle'
LOG_DIR = FLAGS.log_dir
if not os.path.exists(LOG_DIR): os.mkdir(LOG_DIR)
LOG_FOUT = open(os.path.join(LOG_DIR, 'log_train.txt'), 'w')
LOG_FOUT.write(str(FLAGS)+'\n')
#Load dictionary of training queries
print ('pickle reading.')
TRAINING_QUERIES= get_queries_dict(TRAIN_FILE)
print ('train finished.')
TEST_QUERIES= get_queries_dict(TEST_FILE)
BN_INIT_DECAY = 0.5
BN_DECAY_DECAY_RATE = 0.5
BN_DECAY_DECAY_STEP = float(DECAY_STEP)
BN_DECAY_CLIP = 0.99
global HARD_NEGATIVES
HARD_NEGATIVES={}
global TRAINING_LATENT_VECTORS
TRAINING_LATENT_VECTORS=[]
def get_bn_decay(batch):
bn_momentum = tf.train.exponential_decay(
BN_INIT_DECAY,
batch*BATCH_NUM_QUERIES,
BN_DECAY_DECAY_STEP,
BN_DECAY_DECAY_RATE,
staircase=True)
bn_decay = tf.minimum(BN_DECAY_CLIP, 1 - bn_momentum)
return bn_decay
def log_string(out_str):
LOG_FOUT.write(out_str+'\n')
LOG_FOUT.flush()
print(out_str)
#learning rate halfed every 5 epoch
def get_learning_rate(epoch):
learning_rate = BASE_LEARNING_RATE*((0.9)**(epoch//5))
learning_rate = tf.maximum(learning_rate, 0.00001) # CLIP THE LEARNING RATE!
return learning_rate
def train():
global HARD_NEGATIVES
with tf.Graph().as_default():
with tf.device('/gpu:'+str(GPU_INDEX)):
print("In Graph")
query= placeholder_inputs(BATCH_NUM_QUERIES, 1, NUM_POINTS)
positives= placeholder_inputs(BATCH_NUM_QUERIES, POSITIVES_PER_QUERY, NUM_POINTS)
negatives= placeholder_inputs(BATCH_NUM_QUERIES, NEGATIVES_PER_QUERY, NUM_POINTS)
other_negatives= placeholder_inputs(BATCH_NUM_QUERIES,1, NUM_POINTS)
is_training_pl = tf.placeholder(tf.bool, shape=())
print(is_training_pl)
batch = tf.Variable(0)
epoch_num = tf.placeholder(tf.float32, shape=())
bn_decay = get_bn_decay(batch)
tf.summary.scalar('bn_decay', bn_decay)
with tf.variable_scope("query_triplets") as scope:
vecs= tf.concat([query, positives, negatives, other_negatives],1)
print(vecs)
out_vecs, weights= forward(vecs, is_training_pl, bn_decay=bn_decay)
print(out_vecs)
q_vec, pos_vecs, neg_vecs, other_neg_vec= tf.split(out_vecs, [1,POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY,1],1)
print(q_vec)
print(pos_vecs)
print(neg_vecs)
print(other_neg_vec)
#loss = lazy_triplet_loss(q_vec, pos_vecs, neg_vecs, MARGIN1)
#loss = softmargin_loss(q_vec, pos_vecs, neg_vecs)
#loss = quadruplet_loss(q_vec, pos_vecs, neg_vecs, other_neg_vec, MARGIN1, MARGIN2)
loss = lazy_quadruplet_loss(q_vec, pos_vecs, neg_vecs, other_neg_vec, MARGIN1, MARGIN2)
#loss = SARE_loss(q_vec, pos_vecs, neg_vecs)
tf.summary.scalar('loss', loss)
# Get training operator
learning_rate = get_learning_rate(epoch_num)
tf.summary.scalar('learning_rate', learning_rate)
if OPTIMIZER == 'momentum':
optimizer = tf.train.MomentumOptimizer(learning_rate, momentum=MOMENTUM)
elif OPTIMIZER == 'adam':
optimizer = tf.train.AdamOptimizer(learning_rate)
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies(update_ops):
train_op = optimizer.minimize(loss, global_step=batch)
# Add ops to save and restore all the variables.
saver = tf.train.Saver(max_to_keep=50)
# Create a session
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.95)
config = tf.ConfigProto(gpu_options=gpu_options)
config.gpu_options.allow_growth = True
config.allow_soft_placement = True
config.log_device_placement = False
sess = tf.Session(config=config)
# Add summary writers
merged = tf.summary.merge_all()
train_writer = tf.summary.FileWriter(os.path.join(LOG_DIR, 'train'),
sess.graph)
test_writer = tf.summary.FileWriter(os.path.join(LOG_DIR, 'test'))
# Initialize a new model
init = tf.global_variables_initializer()
sess.run(init)
print("Initialized")
# Restore a model
#saver.restore(sess, os.path.join(LOG_DIR, "model2_21001.ckpt"))
#print("Model restored.")
ops = {'query': query,
'positives': positives,
'negatives': negatives,
'other_negatives': other_negatives,
'is_training_pl': is_training_pl,
'loss': loss,
'train_op': train_op,
'merged': merged,
'step': batch,
'epoch_num': epoch_num,
'q_vec':q_vec,
'pos_vecs': pos_vecs,
'neg_vecs': neg_vecs,
'other_neg_vec': other_neg_vec}
for epoch in range(MAX_EPOCH):
print(epoch)
print()
log_string('**** EPOCH %03d ****' % (epoch))
sys.stdout.flush()
train_one_epoch(sess, ops, train_writer, test_writer, epoch, saver)
def train_one_epoch(sess, ops, train_writer, test_writer, epoch, saver):
global HARD_NEGATIVES
global TRAINING_LATENT_VECTORS
is_training = True
sampled_neg=4000
#number of hard negatives in the training tuple
#which are taken from the sampled negatives
num_to_take=5
# Shuffle train files
train_file_idxs = np.arange(0, len(TRAINING_QUERIES.keys()))
np.random.shuffle(train_file_idxs)
for i in range(len(train_file_idxs)//BATCH_NUM_QUERIES):
batch_keys= train_file_idxs[i*BATCH_NUM_QUERIES:(i+1)*BATCH_NUM_QUERIES]
q_tuples=[]
faulty_tuple=False
no_other_neg=False
for j in range(BATCH_NUM_QUERIES):
if(len(TRAINING_QUERIES[batch_keys[j]]["positives"])<POSITIVES_PER_QUERY):
faulty_tuple=True
break
#print('len TRAINING_LATENT_VECTORS', len(TRAINING_LATENT_VECTORS))
#no cached feature vectors
if(len(TRAINING_LATENT_VECTORS)==0):
#q_tuples.append(get_query_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_neg=[], other_neg=True))
q_tuples.append(get_rotated_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_neg=[], other_neg=True))
# q_tuples.append(get_jittered_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_neg=[], other_neg=True))
elif(len(HARD_NEGATIVES.keys())==0):
query=get_feature_representation(TRAINING_QUERIES[batch_keys[j]]['query'], sess, ops)
random.shuffle(TRAINING_QUERIES[batch_keys[j]]['negatives'])
negatives=TRAINING_QUERIES[batch_keys[j]]['negatives'][0:sampled_neg]
hard_negs= get_random_hard_negatives(query, negatives, num_to_take)
print('hard_negs', hard_negs)
#q_tuples.append(get_query_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_negs, other_neg=True))
q_tuples.append(get_rotated_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_negs, other_neg=True))
# q_tuples.append(get_jittered_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_negs, other_neg=True))
else:
query=get_feature_representation(TRAINING_QUERIES[batch_keys[j]]['query'], sess, ops)
random.shuffle(TRAINING_QUERIES[batch_keys[j]]['negatives'])
negatives=TRAINING_QUERIES[batch_keys[j]]['negatives'][0:sampled_neg]
hard_negs= get_random_hard_negatives(query, negatives, num_to_take)
hard_negs= list(set().union(HARD_NEGATIVES[batch_keys[j]], hard_negs))
print('hard',hard_negs)
#q_tuples.append(get_query_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_negs, other_neg=True))
q_tuples.append(get_rotated_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_negs, other_neg=True))
# q_tuples.append(get_jittered_tuple(TRAINING_QUERIES[batch_keys[j]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TRAINING_QUERIES, hard_negs, other_neg=True))
if(q_tuples[j][3].shape[0]!=NUM_POINTS):
no_other_neg= True
break
#construct query array
if(faulty_tuple):
log_string('----' + str(i) + '-----')
log_string('----' + 'FAULTY TUPLE' + '-----')
continue
if(no_other_neg):
log_string('----' + str(i) + '-----')
log_string('----' + 'NO OTHER NEG' + '-----')
continue
queries=[]
positives=[]
negatives=[]
other_neg=[]
for k in range(len(q_tuples)):
queries.append(q_tuples[k][0])
positives.append(q_tuples[k][1])
negatives.append(q_tuples[k][2])
other_neg.append(q_tuples[k][3])
queries= np.array(queries)
queries= np.expand_dims(queries,axis=1)
other_neg= np.array(other_neg)
other_neg= np.expand_dims(other_neg,axis=1)
positives= np.array(positives)
negatives= np.array(negatives)
#print('negtives shape', negatives.shape)
log_string('----' + str(i) + '-----')
if(len(queries.shape)!=4):
log_string('----' + 'FAULTY QUERY' + '-----')
continue
feed_dict={ops['query']:queries, ops['positives']:positives, ops['negatives']:negatives, ops['other_negatives']:other_neg, ops['is_training_pl']:is_training, ops['epoch_num']:epoch}
summary, step, train, loss_val = sess.run([ops['merged'], ops['step'],
ops['train_op'], ops['loss']], feed_dict=feed_dict)
train_writer.add_summary(summary, step)
log_string('batch loss: %f' % loss_val)
if(i%200==7):
test_file_idxs = np.arange(0,len(TEST_QUERIES.keys()))
np.random.shuffle(test_file_idxs)
eval_loss=0
eval_batches=5
eval_batches_counted=0
for eval_batch in range(eval_batches):
eval_keys= test_file_idxs[eval_batch*BATCH_NUM_QUERIES:(eval_batch+1)*BATCH_NUM_QUERIES]
eval_tuples=[]
faulty_eval_tuple=False
no_other_neg= False
for e_tup in range(BATCH_NUM_QUERIES):
if(len(TEST_QUERIES[eval_keys[e_tup]]["positives"])<POSITIVES_PER_QUERY):
faulty_eval_tuple=True
break
eval_tuples.append(get_query_tuple(TEST_QUERIES[eval_keys[e_tup]],POSITIVES_PER_QUERY,NEGATIVES_PER_QUERY, TEST_QUERIES, hard_neg=[], other_neg=True))
if(eval_tuples[e_tup][3].shape[0]!=NUM_POINTS):
no_other_neg= True
break
if(faulty_eval_tuple):
log_string('----' + 'FAULTY EVAL TUPLE' + '-----')
continue
if(no_other_neg):
log_string('----' + str(i) + '-----')
log_string('----' + 'NO OTHER NEG EVAL' + '-----')
continue
eval_batches_counted+=1
eval_queries=[]
eval_positives=[]
eval_negatives=[]
eval_other_neg=[]
for tup in range(len(eval_tuples)):
eval_queries.append(eval_tuples[tup][0])
eval_positives.append(eval_tuples[tup][1])
eval_negatives.append(eval_tuples[tup][2])
eval_other_neg.append(eval_tuples[tup][3])
eval_queries= np.array(eval_queries)
eval_queries= np.expand_dims(eval_queries,axis=1)
eval_other_neg= np.array(eval_other_neg)
eval_other_neg= np.expand_dims(eval_other_neg,axis=1)
eval_positives= np.array(eval_positives)
eval_negatives= np.array(eval_negatives)
feed_dict={ops['query']:eval_queries, ops['positives']:eval_positives, ops['negatives']:eval_negatives, ops['other_negatives']:eval_other_neg, ops['is_training_pl']:False, ops['epoch_num']:epoch}
e_summary, e_step, e_loss= sess.run([ops['merged'], ops['step'], ops['loss']], feed_dict=feed_dict)
eval_loss+=e_loss
if(eval_batch==4):
test_writer.add_summary(e_summary, e_step)
average_eval_loss= float(eval_loss)/eval_batches_counted
log_string('\t\t\tEVAL')
log_string('\t\t\teval_loss: %f' %average_eval_loss)
if(epoch>5 and i%10000 ==29):
#update cached feature vectors
TRAINING_LATENT_VECTORS=get_latent_vectors(sess, ops, TRAINING_QUERIES)
print("Updated cached feature vectors")
if(i%1500==1):
modelname = 'model'+str(epoch)+'_'+str(i)+'.ckpt'
save_path = saver.save(sess, os.path.join(LOG_DIR, modelname))
log_string("Model saved in file: %s" % save_path)
def get_feature_representation(filename, sess, ops):
is_training=False
queries=load_pc_files([filename])
queries= np.expand_dims(queries,axis=1)
if(BATCH_NUM_QUERIES-1>0):
fake_queries=np.zeros((BATCH_NUM_QUERIES-1,1,NUM_POINTS,3))
q=np.vstack((queries,fake_queries))
else:
q=queries
fake_pos=np.zeros((BATCH_NUM_QUERIES,POSITIVES_PER_QUERY,NUM_POINTS,3))
fake_neg=np.zeros((BATCH_NUM_QUERIES,NEGATIVES_PER_QUERY,NUM_POINTS,3))
fake_other_neg=np.zeros((BATCH_NUM_QUERIES,1,NUM_POINTS,3))
feed_dict={ops['query']:q, ops['positives']:fake_pos, ops['negatives']:fake_neg, ops['other_negatives']: fake_other_neg, ops['is_training_pl']:is_training}
output=sess.run(ops['q_vec'], feed_dict=feed_dict)
output=output[0]
output=np.squeeze(output)
return output
def get_random_hard_negatives(query_vec, random_negs, num_to_take):
global TRAINING_LATENT_VECTORS
latent_vecs=[]
for j in range(len(random_negs)):
latent_vecs.append(TRAINING_LATENT_VECTORS[random_negs[j]])
latent_vecs=np.array(latent_vecs)
nbrs = KDTree(latent_vecs)
distances, indices = nbrs.query(np.array([query_vec]),k=num_to_take)
hard_negs=np.squeeze(np.array(random_negs)[indices[0]])
hard_negs= hard_negs.tolist()
return hard_negs
def get_latent_vectors(sess, ops, dict_to_process):
is_training=False
train_file_idxs = np.arange(0, len(dict_to_process.keys()))
batch_num= BATCH_NUM_QUERIES*(1+POSITIVES_PER_QUERY+NEGATIVES_PER_QUERY+1)
q_output = []
for q_index in range(len(train_file_idxs)//batch_num):
file_indices=train_file_idxs[q_index*batch_num:(q_index+1)*(batch_num)]
file_names=[]
for index in file_indices:
file_names.append(dict_to_process[index]["query"])
queries=load_pc_files(file_names)
q1=queries[0:BATCH_NUM_QUERIES]
q1=np.expand_dims(q1,axis=1)
q2=queries[BATCH_NUM_QUERIES:BATCH_NUM_QUERIES*(POSITIVES_PER_QUERY+1)]
q2=np.reshape(q2,(BATCH_NUM_QUERIES,POSITIVES_PER_QUERY,NUM_POINTS,3))
q3=queries[BATCH_NUM_QUERIES*(POSITIVES_PER_QUERY+1):BATCH_NUM_QUERIES*(NEGATIVES_PER_QUERY+POSITIVES_PER_QUERY+1)]
q3=np.reshape(q3,(BATCH_NUM_QUERIES,NEGATIVES_PER_QUERY,NUM_POINTS,3))
q4=queries[BATCH_NUM_QUERIES*(NEGATIVES_PER_QUERY+POSITIVES_PER_QUERY+1):BATCH_NUM_QUERIES*(NEGATIVES_PER_QUERY+POSITIVES_PER_QUERY+2)]
q4=np.expand_dims(q4,axis=1)
feed_dict={ops['query']:q1, ops['positives']:q2, ops['negatives']:q3,ops['other_negatives']:q4, ops['is_training_pl']:is_training}
o1, o2, o3, o4=sess.run([ops['q_vec'], ops['pos_vecs'], ops['neg_vecs'], ops['other_neg_vec']], feed_dict=feed_dict)
o1=np.reshape(o1,(-1,o1.shape[-1]))
o2=np.reshape(o2,(-1,o2.shape[-1]))
o3=np.reshape(o3,(-1,o3.shape[-1]))
o4=np.reshape(o4,(-1,o4.shape[-1]))
out=np.vstack((o1,o2,o3,o4))
q_output.append(out)
q_output=np.array(q_output)
if(len(q_output)!=0):
q_output=q_output.reshape(-1,q_output.shape[-1])
print('q_output', q_output)
#handle edge case
print('edge case')
for q_index in range((len(train_file_idxs)//batch_num*batch_num),len(dict_to_process.keys())):
index=train_file_idxs[q_index]
queries=load_pc_files([dict_to_process[index]["query"]])
queries= np.expand_dims(queries,axis=1)
if(BATCH_NUM_QUERIES-1>0):
fake_queries=np.zeros((BATCH_NUM_QUERIES-1,1,NUM_POINTS,3))
q=np.vstack((queries,fake_queries))
else:
q=queries
fake_pos=np.zeros((BATCH_NUM_QUERIES,POSITIVES_PER_QUERY,NUM_POINTS,3))
fake_neg=np.zeros((BATCH_NUM_QUERIES,NEGATIVES_PER_QUERY,NUM_POINTS,3))
fake_other_neg=np.zeros((BATCH_NUM_QUERIES,1,NUM_POINTS,3))
feed_dict={ops['query']:q, ops['positives']:fake_pos, ops['negatives']:fake_neg, ops['other_negatives']:fake_other_neg, ops['is_training_pl']:is_training}
output=sess.run(ops['q_vec'], feed_dict=feed_dict)
output=output[0]
output=np.squeeze(output)
if (q_output.shape[0]!=0):
q_output=np.vstack((q_output,output))
else:
q_output=output
print(q_output.shape)
return q_output
if __name__ == "__main__":
train()