File size: 23,348 Bytes
c985ba4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved
"""
Misc functions, including distributed helpers.

Mostly copy-paste from torchvision references.
"""
import colorsys
import datetime
import functools
import io
import json
import os
import pickle
import subprocess
import time
from collections import OrderedDict, defaultdict, deque
from typing import List, Optional

import numpy as np
import torch
import torch.distributed as dist

# needed due to empty tensor bug in pytorch and torchvision 0.5
import torchvision
from torch import Tensor

__torchvision_need_compat_flag = float(torchvision.__version__.split(".")[1]) < 7
if __torchvision_need_compat_flag:
    from torchvision.ops import _new_empty_tensor
    from torchvision.ops.misc import _output_size


class SmoothedValue(object):
    """Track a series of values and provide access to smoothed values over a
    window or the global series average.
    """

    def __init__(self, window_size=20, fmt=None):
        if fmt is None:
            fmt = "{median:.4f} ({global_avg:.4f})"
        self.deque = deque(maxlen=window_size)
        self.total = 0.0
        self.count = 0
        self.fmt = fmt

    def update(self, value, n=1):
        self.deque.append(value)
        self.count += n
        self.total += value * n

    def synchronize_between_processes(self):
        """
        Warning: does not synchronize the deque!
        """
        if not is_dist_avail_and_initialized():
            return
        t = torch.tensor([self.count, self.total], dtype=torch.float64, device="cuda")
        dist.barrier()
        dist.all_reduce(t)
        t = t.tolist()
        self.count = int(t[0])
        self.total = t[1]

    @property
    def median(self):
        d = torch.tensor(list(self.deque))
        if d.shape[0] == 0:
            return 0
        return d.median().item()

    @property
    def avg(self):
        d = torch.tensor(list(self.deque), dtype=torch.float32)
        return d.mean().item()

    @property
    def global_avg(self):
        if os.environ.get("SHILONG_AMP", None) == "1":
            eps = 1e-4
        else:
            eps = 1e-6
        return self.total / (self.count + eps)

    @property
    def max(self):
        return max(self.deque)

    @property
    def value(self):
        return self.deque[-1]

    def __str__(self):
        return self.fmt.format(
            median=self.median,
            avg=self.avg,
            global_avg=self.global_avg,
            max=self.max,
            value=self.value,
        )


@functools.lru_cache()
def _get_global_gloo_group():
    """
    Return a process group based on gloo backend, containing all the ranks
    The result is cached.
    """

    if dist.get_backend() == "nccl":
        return dist.new_group(backend="gloo")

    return dist.group.WORLD


def all_gather_cpu(data):
    """
    Run all_gather on arbitrary picklable data (not necessarily tensors)
    Args:
        data: any picklable object
    Returns:
        list[data]: list of data gathered from each rank
    """

    world_size = get_world_size()
    if world_size == 1:
        return [data]

    cpu_group = _get_global_gloo_group()

    buffer = io.BytesIO()
    torch.save(data, buffer)
    data_view = buffer.getbuffer()
    device = "cuda" if cpu_group is None else "cpu"
    tensor = torch.ByteTensor(data_view).to(device)

    # obtain Tensor size of each rank
    local_size = torch.tensor([tensor.numel()], device=device, dtype=torch.long)
    size_list = [torch.tensor([0], device=device, dtype=torch.long) for _ in range(world_size)]
    if cpu_group is None:
        dist.all_gather(size_list, local_size)
    else:
        print("gathering on cpu")
        dist.all_gather(size_list, local_size, group=cpu_group)
    size_list = [int(size.item()) for size in size_list]
    max_size = max(size_list)
    assert isinstance(local_size.item(), int)
    local_size = int(local_size.item())

    # receiving Tensor from all ranks
    # we pad the tensor because torch all_gather does not support
    # gathering tensors of different shapes
    tensor_list = []
    for _ in size_list:
        tensor_list.append(torch.empty((max_size,), dtype=torch.uint8, device=device))
    if local_size != max_size:
        padding = torch.empty(size=(max_size - local_size,), dtype=torch.uint8, device=device)
        tensor = torch.cat((tensor, padding), dim=0)
    if cpu_group is None:
        dist.all_gather(tensor_list, tensor)
    else:
        dist.all_gather(tensor_list, tensor, group=cpu_group)

    data_list = []
    for size, tensor in zip(size_list, tensor_list):
        tensor = torch.split(tensor, [size, max_size - size], dim=0)[0]
        buffer = io.BytesIO(tensor.cpu().numpy())
        obj = torch.load(buffer)
        data_list.append(obj)

    return data_list


def all_gather(data):
    """
    Run all_gather on arbitrary picklable data (not necessarily tensors)
    Args:
        data: any picklable object
    Returns:
        list[data]: list of data gathered from each rank
    """

    if os.getenv("CPU_REDUCE") == "1":
        return all_gather_cpu(data)

    world_size = get_world_size()
    if world_size == 1:
        return [data]

    # serialized to a Tensor
    buffer = pickle.dumps(data)
    storage = torch.ByteStorage.from_buffer(buffer)
    tensor = torch.ByteTensor(storage).to("cuda")

    # obtain Tensor size of each rank
    local_size = torch.tensor([tensor.numel()], device="cuda")
    size_list = [torch.tensor([0], device="cuda") for _ in range(world_size)]
    dist.all_gather(size_list, local_size)
    size_list = [int(size.item()) for size in size_list]
    max_size = max(size_list)

    # receiving Tensor from all ranks
    # we pad the tensor because torch all_gather does not support
    # gathering tensors of different shapes
    tensor_list = []
    for _ in size_list:
        tensor_list.append(torch.empty((max_size,), dtype=torch.uint8, device="cuda"))
    if local_size != max_size:
        padding = torch.empty(size=(max_size - local_size,), dtype=torch.uint8, device="cuda")
        tensor = torch.cat((tensor, padding), dim=0)
    dist.all_gather(tensor_list, tensor)

    data_list = []
    for size, tensor in zip(size_list, tensor_list):
        buffer = tensor.cpu().numpy().tobytes()[:size]
        data_list.append(pickle.loads(buffer))

    return data_list


def reduce_dict(input_dict, average=True):
    """
    Args:
        input_dict (dict): all the values will be reduced
        average (bool): whether to do average or sum
    Reduce the values in the dictionary from all processes so that all processes
    have the averaged results. Returns a dict with the same fields as
    input_dict, after reduction.
    """
    world_size = get_world_size()
    if world_size < 2:
        return input_dict
    with torch.no_grad():
        names = []
        values = []
        # sort the keys so that they are consistent across processes
        for k in sorted(input_dict.keys()):
            names.append(k)
            values.append(input_dict[k])
        values = torch.stack(values, dim=0)
        dist.all_reduce(values)
        if average:
            values /= world_size
        reduced_dict = {k: v for k, v in zip(names, values)}
    return reduced_dict


class MetricLogger(object):
    def __init__(self, delimiter="\t"):
        self.meters = defaultdict(SmoothedValue)
        self.delimiter = delimiter

    def update(self, **kwargs):
        for k, v in kwargs.items():
            if isinstance(v, torch.Tensor):
                v = v.item()
            assert isinstance(v, (float, int))
            self.meters[k].update(v)

    def __getattr__(self, attr):
        if attr in self.meters:
            return self.meters[attr]
        if attr in self.__dict__:
            return self.__dict__[attr]
        raise AttributeError("'{}' object has no attribute '{}'".format(type(self).__name__, attr))

    def __str__(self):
        loss_str = []
        for name, meter in self.meters.items():
            # print(name, str(meter))
            # import ipdb;ipdb.set_trace()
            if meter.count > 0:
                loss_str.append("{}: {}".format(name, str(meter)))
        return self.delimiter.join(loss_str)

    def synchronize_between_processes(self):
        for meter in self.meters.values():
            meter.synchronize_between_processes()

    def add_meter(self, name, meter):
        self.meters[name] = meter

    def log_every(self, iterable, print_freq, header=None, logger=None):
        if logger is None:
            print_func = print
        else:
            print_func = logger.info

        i = 0
        if not header:
            header = ""
        start_time = time.time()
        end = time.time()
        iter_time = SmoothedValue(fmt="{avg:.4f}")
        data_time = SmoothedValue(fmt="{avg:.4f}")
        space_fmt = ":" + str(len(str(len(iterable)))) + "d"
        if torch.cuda.is_available():
            log_msg = self.delimiter.join(
                [
                    header,
                    "[{0" + space_fmt + "}/{1}]",
                    "eta: {eta}",
                    "{meters}",
                    "time: {time}",
                    "data: {data}",
                    "max mem: {memory:.0f}",
                ]
            )
        else:
            log_msg = self.delimiter.join(
                [
                    header,
                    "[{0" + space_fmt + "}/{1}]",
                    "eta: {eta}",
                    "{meters}",
                    "time: {time}",
                    "data: {data}",
                ]
            )
        MB = 1024.0 * 1024.0
        for obj in iterable:
            data_time.update(time.time() - end)
            yield obj
            # import ipdb; ipdb.set_trace()
            iter_time.update(time.time() - end)
            if i % print_freq == 0 or i == len(iterable) - 1:
                eta_seconds = iter_time.global_avg * (len(iterable) - i)
                eta_string = str(datetime.timedelta(seconds=int(eta_seconds)))
                if torch.cuda.is_available():
                    print_func(
                        log_msg.format(
                            i,
                            len(iterable),
                            eta=eta_string,
                            meters=str(self),
                            time=str(iter_time),
                            data=str(data_time),
                            memory=torch.cuda.max_memory_allocated() / MB,
                        )
                    )
                else:
                    print_func(
                        log_msg.format(
                            i,
                            len(iterable),
                            eta=eta_string,
                            meters=str(self),
                            time=str(iter_time),
                            data=str(data_time),
                        )
                    )
            i += 1
            end = time.time()
        total_time = time.time() - start_time
        total_time_str = str(datetime.timedelta(seconds=int(total_time)))
        print_func(
            "{} Total time: {} ({:.4f} s / it)".format(
                header, total_time_str, total_time / len(iterable)
            )
        )


def get_sha():
    cwd = os.path.dirname(os.path.abspath(__file__))

    def _run(command):
        return subprocess.check_output(command, cwd=cwd).decode("ascii").strip()

    sha = "N/A"
    diff = "clean"
    branch = "N/A"
    try:
        sha = _run(["git", "rev-parse", "HEAD"])
        subprocess.check_output(["git", "diff"], cwd=cwd)
        diff = _run(["git", "diff-index", "HEAD"])
        diff = "has uncommited changes" if diff else "clean"
        branch = _run(["git", "rev-parse", "--abbrev-ref", "HEAD"])
    except Exception:
        pass
    message = f"sha: {sha}, status: {diff}, branch: {branch}"
    return message


def collate_fn(batch):
    # import ipdb; ipdb.set_trace()
    batch = list(zip(*batch))
    batch[0] = nested_tensor_from_tensor_list(batch[0])
    return tuple(batch)


def _max_by_axis(the_list):
    # type: (List[List[int]]) -> List[int]
    maxes = the_list[0]
    for sublist in the_list[1:]:
        for index, item in enumerate(sublist):
            maxes[index] = max(maxes[index], item)
    return maxes


class NestedTensor(object):
    def __init__(self, tensors, mask: Optional[Tensor]):
        self.tensors = tensors
        self.mask = mask
        if mask == "auto":
            self.mask = torch.zeros_like(tensors).to(tensors.device)
            if self.mask.dim() == 3:
                self.mask = self.mask.sum(0).to(bool)
            elif self.mask.dim() == 4:
                self.mask = self.mask.sum(1).to(bool)
            else:
                raise ValueError(
                    "tensors dim must be 3 or 4 but {}({})".format(
                        self.tensors.dim(), self.tensors.shape
                    )
                )

    def imgsize(self):
        res = []
        for i in range(self.tensors.shape[0]):
            mask = self.mask[i]
            maxH = (~mask).sum(0).max()
            maxW = (~mask).sum(1).max()
            res.append(torch.Tensor([maxH, maxW]))
        return res

    def to(self, device):
        # type: (Device) -> NestedTensor # noqa
        cast_tensor = self.tensors.to(device)
        mask = self.mask
        if mask is not None:
            assert mask is not None
            cast_mask = mask.to(device)
        else:
            cast_mask = None
        return NestedTensor(cast_tensor, cast_mask)

    def to_img_list_single(self, tensor, mask):
        assert tensor.dim() == 3, "dim of tensor should be 3 but {}".format(tensor.dim())
        maxH = (~mask).sum(0).max()
        maxW = (~mask).sum(1).max()
        img = tensor[:, :maxH, :maxW]
        return img

    def to_img_list(self):
        """remove the padding and convert to img list

        Returns:
            [type]: [description]
        """
        if self.tensors.dim() == 3:
            return self.to_img_list_single(self.tensors, self.mask)
        else:
            res = []
            for i in range(self.tensors.shape[0]):
                tensor_i = self.tensors[i]
                mask_i = self.mask[i]
                res.append(self.to_img_list_single(tensor_i, mask_i))
            return res

    @property
    def device(self):
        return self.tensors.device

    def decompose(self):
        return self.tensors, self.mask

    def __repr__(self):
        return str(self.tensors)

    @property
    def shape(self):
        return {"tensors.shape": self.tensors.shape, "mask.shape": self.mask.shape}


def nested_tensor_from_tensor_list(tensor_list: List[Tensor]):
    # TODO make this more general
    if tensor_list[0].ndim == 3:
        if torchvision._is_tracing():
            # nested_tensor_from_tensor_list() does not export well to ONNX
            # call _onnx_nested_tensor_from_tensor_list() instead
            return _onnx_nested_tensor_from_tensor_list(tensor_list)

        # TODO make it support different-sized images
        max_size = _max_by_axis([list(img.shape) for img in tensor_list])
        # min_size = tuple(min(s) for s in zip(*[img.shape for img in tensor_list]))
        batch_shape = [len(tensor_list)] + max_size
        b, c, h, w = batch_shape
        dtype = tensor_list[0].dtype
        device = tensor_list[0].device
        tensor = torch.zeros(batch_shape, dtype=dtype, device=device)
        mask = torch.ones((b, h, w), dtype=torch.bool, device=device)
        for img, pad_img, m in zip(tensor_list, tensor, mask):
            pad_img[: img.shape[0], : img.shape[1], : img.shape[2]].copy_(img)
            m[: img.shape[1], : img.shape[2]] = False
    else:
        raise ValueError("not supported")
    return NestedTensor(tensor, mask)


# _onnx_nested_tensor_from_tensor_list() is an implementation of
# nested_tensor_from_tensor_list() that is supported by ONNX tracing.
@torch.jit.unused
def _onnx_nested_tensor_from_tensor_list(tensor_list: List[Tensor]) -> NestedTensor:
    max_size = []
    for i in range(tensor_list[0].dim()):
        max_size_i = torch.max(
            torch.stack([img.shape[i] for img in tensor_list]).to(torch.float32)
        ).to(torch.int64)
        max_size.append(max_size_i)
    max_size = tuple(max_size)

    # work around for
    # pad_img[: img.shape[0], : img.shape[1], : img.shape[2]].copy_(img)
    # m[: img.shape[1], :img.shape[2]] = False
    # which is not yet supported in onnx
    padded_imgs = []
    padded_masks = []
    for img in tensor_list:
        padding = [(s1 - s2) for s1, s2 in zip(max_size, tuple(img.shape))]
        padded_img = torch.nn.functional.pad(img, (0, padding[2], 0, padding[1], 0, padding[0]))
        padded_imgs.append(padded_img)

        m = torch.zeros_like(img[0], dtype=torch.int, device=img.device)
        padded_mask = torch.nn.functional.pad(m, (0, padding[2], 0, padding[1]), "constant", 1)
        padded_masks.append(padded_mask.to(torch.bool))

    tensor = torch.stack(padded_imgs)
    mask = torch.stack(padded_masks)

    return NestedTensor(tensor, mask=mask)


def setup_for_distributed(is_master):
    """
    This function disables printing when not in master process
    """
    import builtins as __builtin__

    builtin_print = __builtin__.print

    def print(*args, **kwargs):
        force = kwargs.pop("force", False)
        if is_master or force:
            builtin_print(*args, **kwargs)

    __builtin__.print = print


def is_dist_avail_and_initialized():
    if not dist.is_available():
        return False
    if not dist.is_initialized():
        return False
    return True


def get_world_size():
    if not is_dist_avail_and_initialized():
        return 1
    return dist.get_world_size()


def get_rank():
    if not is_dist_avail_and_initialized():
        return 0
    return dist.get_rank()


def is_main_process():
    return get_rank() == 0


def save_on_master(*args, **kwargs):
    if is_main_process():
        torch.save(*args, **kwargs)


def init_distributed_mode(args):
    if "WORLD_SIZE" in os.environ and os.environ["WORLD_SIZE"] != "":  # 'RANK' in os.environ and
        args.rank = int(os.environ["RANK"])
        args.world_size = int(os.environ["WORLD_SIZE"])
        args.gpu = args.local_rank = int(os.environ["LOCAL_RANK"])

        # launch by torch.distributed.launch
        # Single node
        #   python -m torch.distributed.launch --nproc_per_node=8 main.py --world-size 1 --rank 0 ...
        # Multi nodes
        #   python -m torch.distributed.launch --nproc_per_node=8 main.py --world-size 2 --rank 0 --dist-url 'tcp://IP_OF_NODE0:FREEPORT' ...
        #   python -m torch.distributed.launch --nproc_per_node=8 main.py --world-size 2 --rank 1 --dist-url 'tcp://IP_OF_NODE0:FREEPORT' ...
        # args.rank = int(os.environ.get('OMPI_COMM_WORLD_RANK'))
        # local_world_size = int(os.environ['GPU_PER_NODE_COUNT'])
        # args.world_size = args.world_size * local_world_size
        # args.gpu = args.local_rank = int(os.environ['LOCAL_RANK'])
        # args.rank = args.rank * local_world_size + args.local_rank
        print(
            "world size: {}, rank: {}, local rank: {}".format(
                args.world_size, args.rank, args.local_rank
            )
        )
        print(json.dumps(dict(os.environ), indent=2))
    elif "SLURM_PROCID" in os.environ:
        args.rank = int(os.environ["SLURM_PROCID"])
        args.gpu = args.local_rank = int(os.environ["SLURM_LOCALID"])
        args.world_size = int(os.environ["SLURM_NPROCS"])

        print(
            "world size: {}, world rank: {}, local rank: {}, device_count: {}".format(
                args.world_size, args.rank, args.local_rank, torch.cuda.device_count()
            )
        )
    else:
        print("Not using distributed mode")
        args.distributed = False
        args.world_size = 1
        args.rank = 0
        args.local_rank = 0
        return

    print("world_size:{} rank:{} local_rank:{}".format(args.world_size, args.rank, args.local_rank))
    args.distributed = True
    torch.cuda.set_device(args.local_rank)
    args.dist_backend = "nccl"
    print("| distributed init (rank {}): {}".format(args.rank, args.dist_url), flush=True)

    torch.distributed.init_process_group(
        backend=args.dist_backend,
        world_size=args.world_size,
        rank=args.rank,
        init_method=args.dist_url,
    )

    print("Before torch.distributed.barrier()")
    torch.distributed.barrier()
    print("End torch.distributed.barrier()")
    setup_for_distributed(args.rank == 0)


@torch.no_grad()
def accuracy(output, target, topk=(1,)):
    """Computes the precision@k for the specified values of k"""
    if target.numel() == 0:
        return [torch.zeros([], device=output.device)]
    maxk = max(topk)
    batch_size = target.size(0)

    _, pred = output.topk(maxk, 1, True, True)
    pred = pred.t()
    correct = pred.eq(target.view(1, -1).expand_as(pred))

    res = []
    for k in topk:
        correct_k = correct[:k].view(-1).float().sum(0)
        res.append(correct_k.mul_(100.0 / batch_size))
    return res


@torch.no_grad()
def accuracy_onehot(pred, gt):
    """_summary_

    Args:
        pred (_type_): n, c
        gt (_type_): n, c
    """
    tp = ((pred - gt).abs().sum(-1) < 1e-4).float().sum()
    acc = tp / gt.shape[0] * 100
    return acc


def interpolate(input, size=None, scale_factor=None, mode="nearest", align_corners=None):
    # type: (Tensor, Optional[List[int]], Optional[float], str, Optional[bool]) -> Tensor
    """
    Equivalent to nn.functional.interpolate, but with support for empty batch sizes.
    This will eventually be supported natively by PyTorch, and this
    class can go away.
    """
    if __torchvision_need_compat_flag < 0.7:
        if input.numel() > 0:
            return torch.nn.functional.interpolate(input, size, scale_factor, mode, align_corners)

        output_shape = _output_size(2, input, size, scale_factor)
        output_shape = list(input.shape[:-2]) + list(output_shape)
        return _new_empty_tensor(input, output_shape)
    else:
        return torchvision.ops.misc.interpolate(input, size, scale_factor, mode, align_corners)


class color_sys:
    def __init__(self, num_colors) -> None:
        self.num_colors = num_colors
        colors = []
        for i in np.arange(0.0, 360.0, 360.0 / num_colors):
            hue = i / 360.0
            lightness = (50 + np.random.rand() * 10) / 100.0
            saturation = (90 + np.random.rand() * 10) / 100.0
            colors.append(
                tuple([int(j * 255) for j in colorsys.hls_to_rgb(hue, lightness, saturation)])
            )
        self.colors = colors

    def __call__(self, idx):
        return self.colors[idx]


def inverse_sigmoid(x, eps=1e-3):
    x = x.clamp(min=0, max=1)
    x1 = x.clamp(min=eps)
    x2 = (1 - x).clamp(min=eps)
    return torch.log(x1 / x2)


def clean_state_dict(state_dict):
    new_state_dict = OrderedDict()
    for k, v in state_dict.items():
        if k[:7] == "module.":
            k = k[7:]  # remove `module.`
        new_state_dict[k] = v
    return new_state_dict