-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathutils.py
489 lines (426 loc) · 22.5 KB
/
utils.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
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
import os
import torch
import torch.distributed as dist
# from torch._six import inf
import argparse
from config import get_config
import torch.nn.functional as F
from time import time
def parse_option():
parser = argparse.ArgumentParser('BAE-ViT training and evaluation script', add_help=False)
parser.add_argument('--cfg', type=str, required=True, metavar="FILE", help='path to config file', )
parser.add_argument(
"--opts",
help="Modify config options by adding 'KEY VALUE' pairs. ",
default=None,
nargs='+',
)
# easy config modification
parser.add_argument('--batch-size', type=int, help="batch size for single GPU")
parser.add_argument('--data-path', type=str, help='path to dataset')
parser.add_argument('--zip', action='store_true', help='use zipped dataset instead of folder dataset')
parser.add_argument('--cache-mode', type=str, default='part', choices=['no', 'full', 'part'],
help='no: no cache, '
'full: cache all data, '
'part: sharding the dataset into nonoverlapping pieces and only cache one piece')
parser.add_argument('--pretrained',
help='pretrained weight from checkpoint, could be imagenet22k pretrained weight')
parser.add_argument('--resume', help='resume from checkpoint')
parser.add_argument('--accumulation-steps', type=int, help="gradient accumulation steps")
parser.add_argument('--use-checkpoint', action='store_true',
help="whether to use gradient checkpointing to save memory")
parser.add_argument('--disable_amp', action='store_true', help='Disable pytorch amp')
parser.add_argument('--amp-opt-level', type=str, choices=['O0', 'O1', 'O2'],
help='mixed precision opt level, if O0, no amp is used (deprecated!)')
parser.add_argument('--output', default='output', type=str, metavar='PATH',
help='root of output folder, the full path is <output>/<model_name>/<tag> (default: output)')
parser.add_argument('--tag', help='tag of experiment')
parser.add_argument('--eval', action='store_true', help='Perform evaluation only')
parser.add_argument('--throughput', action='store_true', help='Test throughput only')
# distributed training
parser.add_argument("--local_rank", type=int, required=True, help='local rank for DistributedDataParallel')
# for acceleration
parser.add_argument('--fused_window_process', action='store_true',
help='Fused window shift & window partition, similar for reversed part.')
parser.add_argument('--fused_layernorm', action='store_true', help='Use fused layernorm.')
## overwrite optimizer in config (*.yaml) if specified, e.g., fused_adam/fused_lamb
parser.add_argument('--optim', type=str,
help='overwrite optimizer if provided, can be adamw/sgd/fused_adam/fused_lamb.')
# for medical imaging
parser.add_argument("--max_save", type=int, default=1)
parser.add_argument("--criterion", type=str, default='l1')
parser.add_argument("--save_preds", action='store_true')
parser.add_argument("--debug", action='store_true')
parser.add_argument("--is_pretrained", action='store_true')
parser.add_argument("--pretrained_path", type=str, default='')
parser.add_argument("--random_seed", type=int, default=0)
parser.add_argument("--reg_resume", type=str, default='')
parser.add_argument("--epochs", type=int, default=-1)
parser.add_argument("--multicrop_test", action='store_true')
parser.add_argument("--multi_model", action='store_true')
parser.add_argument("--model_ckpts", nargs='+', default=None)
#parser.add_argument("--freeze_image_branch", type=str, default='')]
parser.add_argument("--freeze_image_branch", action='store_true')
parser.add_argument("--gender_dim", type=int, default=-1)
parser.add_argument("--alpha", type=float, default=-1.0)
parser.add_argument("--gender_filter", type=int, default=-1)
parser.add_argument("--base_lr", type=float, default=-1.0)
parser.add_argument("--layer_lr_decay", type=float, default=-1.0)
parser.add_argument("--gen_heatmap", action='store_true')
parser.add_argument("--no_test_crop", action='store_true')
parser.add_argument("--test_pad", action='store_true')
parser.add_argument("--warmup", type=int, default=-1)
parser.add_argument("--test_only", action='store_true')
## The following argument is abandoned
parser.add_argument("--task", type=str, default='classification')
args, unparsed = parser.parse_known_args()
config = get_config(args)
return args, config
def get_stats(output, func, logger, return_worst=True):
# func: either max or min
# return_worst: If True, then return the filename with the worst val metric; otherwise the ckpt with the best val metric will be returned
func_name = func.__name__
assert func_name in ['max', 'min'], f"The function ({func_name}) is not supported for obtaining the statistics of all ckpts"
reverse = True if func_name == 'max' else False
checkpoints = os.listdir(output)
checkpoints = [ckpt for ckpt in checkpoints if ckpt.endswith('pth')]
logger.info(f"All checkpoints found in {output}: {checkpoints}")
num_ckpts = len(checkpoints)
val_metric_list = []
for ckpt in checkpoints:
ckpt_path = os.path.join(output, ckpt)
ckpt_value = torch.load(ckpt_path, map_location='cpu')
val_metric = ckpt_value['val_metric']
val_metric_list.append([ckpt_path, val_metric])
if len(val_metric_list):
val_metric_list.sort(reverse=reverse, key=lambda x: x[1])
path, metric = val_metric_list[-1] if return_worst else val_metric_list[0]
else:
path = None
metric = None
return path, metric, num_ckpts
def load_checkpoint_weight_only(path, model, logger):
checkpoint = torch.load(path, map_location='cpu')
msg = model.load_state_dict(checkpoint['model'], strict=False)
#logger.info(msg)
del checkpoint
torch.cuda.empty_cache()
def load_checkpoint(config, model, optimizer, lr_scheduler, loss_scaler, logger):
logger.info(f"==============> Resuming form {config.MODEL.RESUME}....................")
if config.MODEL.RESUME.startswith('https'):
checkpoint = torch.hub.load_state_dict_from_url(
config.MODEL.RESUME, map_location='cpu', check_hash=True)
else:
checkpoint = torch.load(config.MODEL.RESUME, map_location='cpu')
msg = model.load_state_dict(checkpoint['model'], strict=False)
logger.info(msg)
max_accuracy = 0.0
if not config.EVAL_MODE and 'optimizer' in checkpoint and 'lr_scheduler' in checkpoint and 'epoch' in checkpoint:
optimizer.load_state_dict(checkpoint['optimizer'])
lr_scheduler.load_state_dict(checkpoint['lr_scheduler'])
config.defrost()
config.TRAIN.START_EPOCH = checkpoint['epoch'] + 1
config.freeze()
if 'scaler' in checkpoint:
loss_scaler.load_state_dict(checkpoint['scaler'])
logger.info(f"=> loaded successfully '{config.MODEL.RESUME}' (epoch {checkpoint['epoch']})")
# if 'max_accuracy' in checkpoint:
# max_accuracy = checkpoint['max_accuracy']
if 'opt_val_metric' in checkpoint:
opt_val_metric = checkpoint['opt_val_metric']
else:
opt_val_metric = None
del checkpoint
torch.cuda.empty_cache()
return opt_val_metric
def load_reg_checkpoint(config, model, optimizer, lr_scheduler, loss_scaler, logger):
logger.info(f"==============> Resuming form {config.MI.REG_RESUME}....................")
checkpoint = torch.load(config.MI.REG_RESUME, map_location='cpu')
model_state = model.state_dict()
ckpt_state = checkpoint['model']
## Drop head weights
mnames = ['head.weight', 'head.bias']
if mnames[-1] in model_state:
if mnames[-1] in ckpt_state and ckpt_state[mnames[-1]].shape != model_state[mnames[-1]].shape:
for mname in mnames:
ckpt_state.pop(mname)
## Drop attn mask
for key in model_state:
if ('attn_mask' in key or 'relative_position_index' in key) and key in ckpt_state:
ckpt_state.pop(key)
## Interpolate relative pos bias table
all_key_matched = True
for key in model_state:
if 'relative_position_bias_table' in key and key in ckpt_state: ## Swin
# table: (L, num_heads)
l_target, nh_target = model_state[key].size()
l_source, nh_source = ckpt_state[key].size()
if l_target != l_source:
print(f'Interpolate the position bias table from {l_source} to {l_target}...')
assert nh_source == nh_target, f"Different number of heads detected! {nh_source} VS {nh_target}..."
sl_source = int(l_source ** 0.5)
sl_target = int(l_target ** 0.5)
value = F.interpolate(ckpt_state[key].permute(1,0).view(1, nh_source, sl_source, sl_source), size=(sl_target, sl_target), mode='bicubic') # (1, nh, l2, l2)
value = value.reshape(nh_source, l_target).permute(1, 0)
ckpt_state[key] = value
elif 'attention_biases' in key and key in ckpt_state:
nh_target, l_target = model_state[key].size()
nh_source, l_source = ckpt_state[key].size()
if l_target != l_source:
print(f'Interpolate the position bias table from {l_source} to {l_target}...')
assert nh_source == nh_target, f"Different number of heads detected! {nh_source} VS {nh_target}..."
sl_source = int(l_source ** 0.5)
sl_target = int(l_target ** 0.5)
value = F.interpolate(ckpt_state[key].view(1, nh_source, sl_source, sl_source), size=(sl_target, sl_target), mode='bicubic') # (1, nh, l2, l2)
value = value.reshape(nh_source, l_target)
ckpt_state[key] = value
elif 'local_conv.c' in key and key in ckpt_state:
_, _, s1, s2 = model_state[key].size()
_, _, t1, t2 = ckpt_state[key].size()
if (s1 != t1) or (s2 != t2):
print(f'Interpolate the local convolution size ({key}) from ({s1},{s2}) to ({t1},{t2})...')
value = F.interpolate(ckpt_state[key], size=(s1, s2), mode='bicubic')
ckpt_state[key] = value
elif key == 'pos_embed': # DeiT
# pos_embed (1, n, dim)
_, l_target, dim_target = model_state[key].size()
_, l_source, dim_source = ckpt_state[key].size()
if l_target != l_source:
print(f'Interpolate the position bias table from {l_source} to {l_target}...')
assert dim_target == dim_source, f"Different number of heads detected! {dim_source} VS {dim_target}..."
sl_source = int((l_source-1) ** 0.5)
sl_target = int((l_target-1) ** 0.5)
## pos embed for cls token is at the first row
value = F.interpolate(ckpt_state[key][0, 1:, :].permute(1, 0).view(1, dim_source, sl_source, sl_source), size=(sl_target, sl_target), mode='bicubic')
value = value.reshape(dim_source, sl_target*sl_target).permute(1, 0)
value = torch.cat([ckpt_state[key][0, 0, :].unsqueeze(0), value])
ckpt_state[key] = value.unsqueeze(0)
elif key == 'cls_token': # DeiT/ViT
value = ckpt_state[key]
value = torch.stack([value[0], value[0]])
ckpt_state[key] = value
if key in ckpt_state:
model_state[key] = ckpt_state[key]
else:
all_key_matched = False
print(f"Key not found in the checkpoint: {key}")
if all_key_matched:
print(f"All keys successfully matched!")
msg = model.load_state_dict(model_state, strict=False)
logger.info(msg)
del checkpoint
torch.cuda.empty_cache()
def load_pretrained(config, model, logger):
logger.info(f"==============> Loading weight {config.MODEL.PRETRAINED} for fine-tuning......")
checkpoint = torch.load(config.MODEL.PRETRAINED, map_location='cpu')
state_dict = checkpoint['model']
# delete relative_position_index since we always re-init it
relative_position_index_keys = [k for k in state_dict.keys() if "relative_position_index" in k]
for k in relative_position_index_keys:
del state_dict[k]
# delete relative_coords_table since we always re-init it
relative_position_index_keys = [k for k in state_dict.keys() if "relative_coords_table" in k]
for k in relative_position_index_keys:
del state_dict[k]
# delete attn_mask since we always re-init it
attn_mask_keys = [k for k in state_dict.keys() if "attn_mask" in k]
for k in attn_mask_keys:
del state_dict[k]
# bicubic interpolate relative_position_bias_table if not match
relative_position_bias_table_keys = [k for k in state_dict.keys() if "relative_position_bias_table" in k]
for k in relative_position_bias_table_keys:
relative_position_bias_table_pretrained = state_dict[k]
relative_position_bias_table_current = model.state_dict()[k]
L1, nH1 = relative_position_bias_table_pretrained.size()
L2, nH2 = relative_position_bias_table_current.size()
if nH1 != nH2:
logger.warning(f"Error in loading {k}, passing......")
else:
if L1 != L2:
# bicubic interpolate relative_position_bias_table if not match
S1 = int(L1 ** 0.5)
S2 = int(L2 ** 0.5)
relative_position_bias_table_pretrained_resized = torch.nn.functional.interpolate(
relative_position_bias_table_pretrained.permute(1, 0).view(1, nH1, S1, S1), size=(S2, S2),
mode='bicubic')
state_dict[k] = relative_position_bias_table_pretrained_resized.view(nH2, L2).permute(1, 0)
# bicubic interpolate absolute_pos_embed if not match
absolute_pos_embed_keys = [k for k in state_dict.keys() if "absolute_pos_embed" in k]
for k in absolute_pos_embed_keys:
# dpe
absolute_pos_embed_pretrained = state_dict[k]
absolute_pos_embed_current = model.state_dict()[k]
_, L1, C1 = absolute_pos_embed_pretrained.size()
_, L2, C2 = absolute_pos_embed_current.size()
if C1 != C1:
logger.warning(f"Error in loading {k}, passing......")
else:
if L1 != L2:
S1 = int(L1 ** 0.5)
S2 = int(L2 ** 0.5)
absolute_pos_embed_pretrained = absolute_pos_embed_pretrained.reshape(-1, S1, S1, C1)
absolute_pos_embed_pretrained = absolute_pos_embed_pretrained.permute(0, 3, 1, 2)
absolute_pos_embed_pretrained_resized = torch.nn.functional.interpolate(
absolute_pos_embed_pretrained, size=(S2, S2), mode='bicubic')
absolute_pos_embed_pretrained_resized = absolute_pos_embed_pretrained_resized.permute(0, 2, 3, 1)
absolute_pos_embed_pretrained_resized = absolute_pos_embed_pretrained_resized.flatten(1, 2)
state_dict[k] = absolute_pos_embed_pretrained_resized
# check classifier, if not match, then re-init classifier to zero
head_bias_pretrained = state_dict['head.bias']
Nc1 = head_bias_pretrained.shape[0]
Nc2 = model.head.bias.shape[0]
if (Nc1 != Nc2):
if Nc1 == 21841 and Nc2 == 1000:
logger.info("loading ImageNet-22K weight to ImageNet-1K ......")
map22kto1k_path = f'data/map22kto1k.txt'
with open(map22kto1k_path) as f:
map22kto1k = f.readlines()
map22kto1k = [int(id22k.strip()) for id22k in map22kto1k]
state_dict['head.weight'] = state_dict['head.weight'][map22kto1k, :]
state_dict['head.bias'] = state_dict['head.bias'][map22kto1k]
else:
torch.nn.init.constant_(model.head.bias, 0.)
torch.nn.init.constant_(model.head.weight, 0.)
del state_dict['head.weight']
del state_dict['head.bias']
logger.warning(f"Error in loading classifier head, re-init classifier head to 0")
msg = model.load_state_dict(state_dict, strict=False)
logger.warning(msg)
logger.info(f"=> loaded successfully '{config.MODEL.PRETRAINED}'")
del checkpoint
torch.cuda.empty_cache()
def save_checkpoint(config, epoch, model, optimizer, lr_scheduler, loss_scaler, val_loss, val_metric, opt_val_metric, logger):
save_state = {'model': model.state_dict(),
'optimizer': optimizer.state_dict(),
'lr_scheduler': lr_scheduler.state_dict(),
'scaler': loss_scaler.state_dict(),
'epoch': epoch,
'config': config,
'val_loss': val_loss,
'val_metric': val_metric,
'opt_val_metric': opt_val_metric,
}
save_path = os.path.join(config.OUTPUT, f'ckpt_epoch_{epoch}.pth')
logger.info(f"{save_path} saving......")
for _ in range(200):
try:
torch.save(save_state, save_path)
logger.info(f"{save_path} saved !!!")
break
except:
logger.info(f"Save file {save_path} failed! re-try after 3 seconds")
time.sleep(3)
def get_grad_norm(parameters, norm_type=2):
if isinstance(parameters, torch.Tensor):
parameters = [parameters]
parameters = list(filter(lambda p: p.grad is not None, parameters))
norm_type = float(norm_type)
total_norm = 0
for p in parameters:
param_norm = p.grad.data.norm(norm_type)
total_norm += param_norm.item() ** norm_type
total_norm = total_norm ** (1. / norm_type)
return total_norm
def auto_resume_helper(output_dir, logger):
checkpoints = os.listdir(output_dir)
checkpoints = [ckpt for ckpt in checkpoints if ckpt.endswith('pth')]
logger.info(f"All checkpoints founded in {output_dir}: {checkpoints}")
if len(checkpoints) > 0:
#latest_checkpoint = max([os.path.join(output_dir, d) for d in checkpoints], key=os.path.getmtime)
latest_checkpoint = max([os.path.join(output_dir, d) for d in checkpoints], key=lambda x: int(os.path.basename(x).rsplit('_', 1)[1].split('.')[0]))
logger.info(f"The latest checkpoint founded: {latest_checkpoint}")
resume_file = latest_checkpoint
else:
resume_file = None
return resume_file
def reduce_tensor(tensor):
rt = tensor.clone()
dist.all_reduce(rt, op=dist.ReduceOp.SUM)
rt /= dist.get_world_size()
return rt
def ampscaler_get_grad_norm(parameters, norm_type: float = 2.0) -> torch.Tensor:
if isinstance(parameters, torch.Tensor):
parameters = [parameters]
parameters = [p for p in parameters if p.grad is not None]
norm_type = float(norm_type)
if len(parameters) == 0:
return torch.tensor(0.)
device = parameters[0].grad.device
if norm_type == float('inf'):
total_norm = max(p.grad.detach().abs().max().to(device) for p in parameters)
else:
total_norm = torch.norm(torch.stack([torch.norm(p.grad.detach(),
norm_type).to(device) for p in parameters]), norm_type)
return total_norm
def percentage_loss(y_pred, y_truth, alpha=1.0):
l1 = ((y_pred - y_truth)/y_truth).abs().mean()
ld = ((y_pred.mean() - y_truth.mean())/y_truth.mean()).abs()
return l1 + alpha * ld
def ccc_loss(y_pred, y_truth):
mean_truth = y_truth.mean()
mean_pred = y_pred.mean()
cov = ((y_truth - mean_truth) * (y_pred - mean_pred)).mean()
var = y_truth.var() + y_pred.var()
mse = (mean_pred - mean_truth) ** 2
return (2*cov) / (var + mse)
class NativeScalerWithGradNormCount:
state_dict_key = "amp_scaler"
def __init__(self):
self._scaler = torch.cuda.amp.GradScaler()
def __call__(self, loss, optimizer, clip_grad=None, parameters=None, create_graph=False, update_grad=True):
self._scaler.scale(loss).backward(create_graph=create_graph)
if update_grad:
if clip_grad is not None:
assert parameters is not None
self._scaler.unscale_(optimizer) # unscale the gradients of optimizer's assigned params in-place
norm = torch.nn.utils.clip_grad_norm_(parameters, clip_grad)
else:
self._scaler.unscale_(optimizer)
norm = ampscaler_get_grad_norm(parameters)
self._scaler.step(optimizer)
self._scaler.update()
else:
norm = None
return norm
def state_dict(self):
return self._scaler.state_dict()
def load_state_dict(self, state_dict):
self._scaler.load_state_dict(state_dict)
class GatherTensor:
"""Computes and stores the average and current value"""
def __init__(self):
self.reset()
def reset(self):
self.pred_list = []
self.target_list = []
self.all_pred_list = []
self.all_target_list = []
def update(self, pred, target):
self.pred_list.append(pred)
self.target_list.append(target)
def sync(self):
### Support gathering tensors with different sizes
n = dist.get_world_size()
pred_list = torch.cat(self.pred_list)
target_list = torch.cat(self.target_list)
local_size = torch.tensor(len(pred_list), device=pred_list.device)
all_sizes = [torch.tensor(0, device=local_size.device, dtype=local_size.dtype) for _ in range(n)]
dist.all_gather(all_sizes, local_size)
max_size = max(all_sizes)
size_diff = max_size.item() - local_size.item()
if size_diff:
new_shape = list(pred_list.shape)
new_shape[0] = size_diff
padding = torch.zeros(new_shape, device=pred_list.device, dtype=pred_list.dtype)
pred_list = torch.cat([pred_list, padding])
new_shape = list(target_list.shape)
new_shape[0] = size_diff
padding = torch.zeros(new_shape, device=target_list.device, dtype=target_list.dtype)
target_list = torch.cat([target_list, padding])
all_pred_list = [torch.zeros(pred_list.shape, device=pred_list.device, dtype=pred_list.dtype) for _ in range(n)]
all_target_list = [torch.zeros(target_list.shape, device=target_list.device, dtype=target_list.dtype) for _ in range(n)]
dist.all_gather(all_pred_list, pred_list)
dist.all_gather(all_target_list, target_list)
for pred, target, size in zip(all_pred_list, all_target_list, all_sizes):
self.all_pred_list.append(pred[:size])
self.all_target_list.append(target[:size])