utils.py

from collections import deque
import os
from collections import defaultdict, deque
import warnings
from xml.dom import NotSupportedErr
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
import torch
import torch.distributed as dist
from continuum.datasets import CIFAR100, ImageFolderDataset, Food101
from continuum import ClassIncremental
from timm.data import create_transform
from timm.data.constants import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD
from torchvision import transforms
try:
    interpolation = torch.transforms.functional.InterpolationMode.BICUBIC
except:
    interpolation = 3


class SmoothedValue(object):
    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):
        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))
        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):
        return self.total / self.count

    @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)


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 v is None:
                continue
            if isinstance(v, torch.Tensor):
                v = v.item()
            assert isinstance(v, (float, int))
            self.meters[k].update(v)

    def update_dict(self, d):
        for k, v in d.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():
            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


class SoftTarget(nn.Module):

    def __init__(self, T=2):
        super(SoftTarget, self).__init__()
        self.T = T

    def forward(self, out_s, out_t):
        loss = F.kl_div(F.log_softmax(out_s/self.T, dim=1),
                        F.softmax(out_t/self.T, dim=1),
                        reduction='batchmean') * self.T * self.T

        return loss

def init_distributed_mode(args):
    if 'RANK' in os.environ and 'WORLD_SIZE' in os.environ:
        args.rank = int(os.environ["RANK"])
        args.world_size = int(os.environ['WORLD_SIZE'])
        args.gpu = int(os.environ['LOCAL_RANK'])
    else:
        print('Not using distributed mode')
        args.distributed = False
        raise NotSupportedErr("not supported yet!")
        return
    args.distributed = True
    torch.cuda.set_device(args.gpu)
    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, init_method=args.dist_url,
                                         world_size=args.world_size, rank=args.rank)
    torch.distributed.barrier()
    setup_for_distributed(args.rank == 0)


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


def setup_for_distributed(is_master):
    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


class ImageNet1000(ImageFolderDataset):
    def __init__(
            self,
            data_path: str,
            train: bool = True,
            download: bool = False,
    ):
        super().__init__(data_path=data_path, train=train, download=download)

    def get_data(self):
        if self.train:
            self.data_path = os.path.join(self.data_path, "train")
        else:
            self.data_path = os.path.join(self.data_path, "val")
        return super().get_data()


def build_dataset(is_train, apply_da, args):
    transform = build_transform(is_train, apply_da, args)

    if args.data_set.lower() == 'cifar':
        dataset = CIFAR100(args.data_path, train=is_train, download=True)
    elif args.data_set.lower() == 'imagenet1000':
        dataset = ImageNet1000(args.data_path, train=is_train)
    elif args.data_set.lower() == 'food101':
        dataset = Food101(args.data_path, train=is_train, download=True)
    else:
        raise ValueError(f'Unknown dataset {args.data_set}.')

    scenario = ClassIncremental(
        dataset,
        initial_increment=args.num_bases,
        increment=args.increment,
        transformations=transform.transforms,
        class_order=args.class_order
    )
    nb_classes = scenario.nb_classes

    return scenario, nb_classes


def build_transform(is_train, apply_da, args):
    if args.aa == 'none':
        args.aa = None

    with warnings.catch_warnings():
        resize_im = args.input_size > 32
        if is_train and apply_da:
            transform = create_transform(
                input_size=args.input_size,
                is_training=True,
                color_jitter=args.color_jitter,
                auto_augment=args.aa,
                interpolation='bicubic',
                re_prob=args.reprob,
                re_mode=args.remode,
                re_count=args.recount,
            )
            if not resize_im:
                transform.transforms[0] = transforms.RandomCrop(
                    args.input_size, padding=4)
            if args.input_size == 32 and args.data_set == 'CIFAR':
                transform.transforms[-1] = transforms.Normalize(
                    (0.5071, 0.4867, 0.4408), (0.2675, 0.2565, 0.2761))
            return transform

        t = []
        if resize_im:
            size = int((256 / 224) * args.input_size)
            t.append(
                transforms.Resize(size, interpolation=interpolation),
            )
            t.append(transforms.CenterCrop(args.input_size))

        t.append(transforms.ToTensor())
        print(args.input_size == 32 and args.data_set == 'CIFAR')
        if args.input_size == 32 and args.data_set == 'CIFAR':
            t.append(transforms.Normalize(
                (0.5071, 0.4867, 0.4408), (0.2675, 0.2565, 0.2761)))
        else:
            t.append(transforms.Normalize(
                IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD))
        return transforms.Compose(t)