import argparse
import os
import shutil
import time
from fastai.transforms import *
from fastai.dataset import *
from fastai.fp16 import *
from fastai.conv_learner import *
from pathlib import *
from fastai import io
import tarfile
import torch
from torch.autograd import Variable
import torch.nn as nn
import torch.nn.parallel
import torch.backends.cudnn as cudnn
import torch.distributed as dist
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torchvision.transforms as transforms
import torchvision.datasets as datasets
import models
import models.cifar10 as cifar10models
from distributed import DistributedDataParallel as DDP
# print(models.cifar10.__dict__)
model_names = sorted(name for name in models.__dict__
if name.islower() and not name.startswith("__")
and callable(models.__dict__[name]))
cifar10_names = sorted(name for name in cifar10models.__dict__
if name.islower() and not name.startswith("__")
and callable(cifar10models.__dict__[name]))
model_names = cifar10_names + model_names
# print(model_names)
# Example usage: python run_fastai.py /home/paperspace/ILSVRC/Data/CLS-LOC/ -a resnext_50_32x4d --epochs 1 -j 4 -b 64 --fp16
parser = argparse.ArgumentParser(description='PyTorch Cifar10 Training')
parser.add_argument('data', metavar='DIR',
help='path to dataset')
parser.add_argument('--save-dir', type=str, default=Path.home()/'imagenet_training',
help='Directory to save logs and models.')
parser.add_argument('--arch', '-a', metavar='ARCH', default='resnet56',
choices=model_names,
help='model architecture: ' +
' | '.join(model_names) +
' (default: resnet56)')
parser.add_argument('-dp', '--data-parallel', default=False, type=bool, help='Use DataParallel')
parser.add_argument('-j', '--workers', default=7, type=int, metavar='N',
help='number of data loading workers (default: 4)')
parser.add_argument('--epochs', default=1, type=int, metavar='N',
help='number of total epochs to run')
parser.add_argument('--cycle-len', default=95, type=float, metavar='N',
help='Length of cycle to run')
# parser.add_argument('--start-epoch', default=0, type=int, metavar='N',
# help='manual epoch number (useful on restarts)')
parser.add_argument('-b', '--batch-size', default=512, type=int,
metavar='N', help='mini-batch size (default: 512)')
parser.add_argument('--lr', '--learning-rate', default=0.8, type=float,
metavar='LR', help='initial learning rate')
parser.add_argument('--momentum', default=0.9, type=float, metavar='M', help='momentum')
parser.add_argument('--weight-decay', '--wd', default=1e-4, type=float,
metavar='W', help='weight decay (default: 1e-4)')
# parser.add_argument('--print-freq', '-p', default=10, type=int,
# metavar='N', help='print frequency (default: 10)')
# parser.add_argument('--resume', default='', type=str, metavar='PATH',
# help='path to latest checkpoint (default: none)')
# parser.add_argument('-e', '--evaluate', dest='evaluate', action='store_true',
# help='evaluate model on validation set')
parser.add_argument('--pretrained', dest='pretrained', action='store_true', help='use pre-trained model')
parser.add_argument('--fp16', action='store_true', help='Run model fp16 mode.')
parser.add_argument('--use-tta', default=False, type=bool, help='Validate model with TTA at the end of traiing.')
parser.add_argument('--sz', default=32, type=int, help='Size of transformed image.')
# parser.add_argument('--decay-int', default=30, type=int, help='Decay LR by 10 every decay-int epochs')
parser.add_argument('--use-clr', default='10,13.68,0.95,0.85', type=str,
help='div,pct,max_mom,min_mom. Pass in a string delimited by commas. Ex: "20,2,0.95,0.85"')
parser.add_argument('--loss-scale', type=float, default=128,
help='Loss scaling, positive power of 2 values can improve fp16 convergence.')
parser.add_argument('--warmup', action='store_true', help='Do a warm-up epoch first')
parser.add_argument('--prof', dest='prof', action='store_true', help='Only run a few iters for profiling.')
parser.add_argument('--dist-url', default='file://sync.file', type=str,
help='url used to set up distributed training')
parser.add_argument('--dist-backend', default='nccl', type=str, help='distributed backend')
parser.add_argument('--world-size', default=1, type=int,
help='Number of GPUs to use. Can either be manually set ' +
'or automatically set by using \'python -m multiproc\'.')
parser.add_argument('--rank', default=0, type=int,
help='Used for multi-process training. Can either be manually set ' +
'or automatically set by using \'python -m multiproc\'.')
def pad(img, p=4, padding_mode='reflect'):
return Image.fromarray(np.pad(np.asarray(img), ((p, p), (p, p), (0, 0)), padding_mode))
class TorchModelData(ModelData):
def __init__(self, path, sz, trn_dl, val_dl, aug_dl=None):
super().__init__(path, trn_dl, val_dl)
self.aug_dl = aug_dl
self.sz = sz
def download_cifar10(data_path):
# (AS) TODO: put this into the fastai library
def untar_file(file_path, save_path):
if file_path.endswith('.tar.gz') or file_path.endswith('.tgz'):
obj = tarfile.open(file_path)
obj.extractall(save_path)
obj.close()
os.remove(file_path)
cifar_url = 'http://files.fast.ai/data/cifar10.tgz' # faster download
# cifar_url = 'http://pjreddie.com/media/files/cifar.tgz'
io.get_data(cifar_url, args.data+'/cifar10.tgz')
untar_file(data_path+'/cifar10.tgz', data_path)
# Loader expects train and test folders to be outside of cifar10 folder
shutil.move(data_path+'/cifar10/train', data_path)
shutil.move(data_path+'/cifar10/test', data_path)
def torch_loader(data_path, size):
if not os.path.exists(data_path+'/train'): download_cifar10(data_path)
# Data loading code
traindir = os.path.join(data_path, 'train')
valdir = os.path.join(data_path, 'test')
normalize = transforms.Normalize(mean=[0.4914 , 0.48216, 0.44653], std=[0.24703, 0.24349, 0.26159])
tfms = [transforms.ToTensor(), normalize]
scale_size = 40
padding = int((scale_size - size) / 2)
train_tfms = transforms.Compose([
pad, # TODO: use `padding` rather than assuming 4
transforms.RandomCrop(size),
transforms.ColorJitter(.25,.25,.25),
transforms.RandomRotation(2),
transforms.RandomHorizontalFlip(),
] + tfms)
train_dataset = datasets.ImageFolder(traindir, train_tfms)
train_sampler = (torch.utils.data.distributed.DistributedSampler(train_dataset)
if args.distributed else None)
train_loader = torch.utils.data.DataLoader(
train_dataset, batch_size=args.batch_size, shuffle=(train_sampler is None),
num_workers=args.workers, pin_memory=True, sampler=train_sampler)
val_tfms = transforms.Compose(tfms)
val_loader = torch.utils.data.DataLoader(
datasets.ImageFolder(valdir, val_tfms),
batch_size=args.batch_size*2, shuffle=False,
num_workers=args.workers, pin_memory=True)
aug_loader = torch.utils.data.DataLoader(
datasets.ImageFolder(valdir, train_tfms),
batch_size=args.batch_size, shuffle=False,
num_workers=args.workers, pin_memory=True)
train_loader = DataPrefetcher(train_loader)
val_loader = DataPrefetcher(val_loader)
aug_loader = DataPrefetcher(aug_loader)
if args.prof:
train_loader.stop_after = 200
val_loader.stop_after = 0
data = TorchModelData(data_path, args.sz, train_loader, val_loader, aug_loader)
return data, train_sampler
# Seems to speed up training by ~2%
class DataPrefetcher():
def __init__(self, loader, stop_after=None):
self.loader = loader
self.dataset = loader.dataset
self.stream = torch.cuda.Stream()
self.stop_after = stop_after
self.next_input = None
self.next_target = None
def __len__(self):
return len(self.loader)
def preload(self):
try:
self.next_input, self.next_target = next(self.loaditer)
except StopIteration:
self.next_input = None
self.next_target = None
return
with torch.cuda.stream(self.stream):
self.next_input = self.next_input.cuda(async=True)
self.next_target = self.next_target.cuda(async=True)
def __iter__(self):
count = 0
self.loaditer = iter(self.loader)
self.preload()
while self.next_input is not None:
torch.cuda.current_stream().wait_stream(self.stream)
input = self.next_input
target = self.next_target
self.preload()
count += 1
yield input, target
if type(self.stop_after) is int and (count > self.stop_after):
break
def top5(output, target):
"""Computes the precision@k for the specified values of k"""
top5 = 5
batch_size = target.size(0)
_, pred = output.topk(top5, 1, True, True)
pred = pred.t()
correct = pred.eq(target.view(1, -1).expand_as(pred))
correct_k = correct[:top5].view(-1).float().sum(0, keepdim=True)
return correct_k.mul_(1.0 / batch_size)
class ImagenetLoggingCallback(Callback):
def __init__(self, save_path, print_every=50):
super().__init__()
self.save_path=save_path
self.print_every=print_every
def on_train_begin(self):
self.batch = 0
self.epoch = 0
self.f = open(self.save_path, "a", 1)
self.log("\ton_train_begin")
def on_epoch_end(self, metrics):
log_str = f'\tEpoch:{self.epoch}\ttrn_loss:{self.last_loss}'
for (k,v) in zip(['val_loss', 'acc', 'top5', ''], metrics): log_str += f'\t{k}:{v}'
self.log(log_str)
self.epoch += 1
def on_batch_end(self, metrics):
self.last_loss = metrics
self.batch += 1
if self.batch % self.print_every == 0:
self.log(f'Epoch: {self.epoch} Batch: {self.batch} Metrics: {metrics}')
def on_train_end(self):
self.log("\ton_train_end")
self.f.close()
def log(self, string):
self.f.write(time.strftime("%Y-%m-%dT%H:%M:%S")+"\t"+string+"\n")
# Logging + saving models
def save_args(name, save_dir):
if (args.rank != 0) or not args.save_dir: return {}
log_dir = f'{save_dir}/training_logs'
os.makedirs(log_dir, exist_ok=True)
return {
'best_save_name': f'{name}_best_model',
'cycle_save_name': f'{name}',
'callbacks': [
ImagenetLoggingCallback(f'{log_dir}/{name}_log.txt')
]
}
def save_sched(sched, save_dir):
if (args.rank != 0) or not args.save_dir: return {}
log_dir = f'{save_dir}/training_logs'
sched.save_path = log_dir
sched.plot_loss()
sched.plot_lr()
def update_model_dir(learner, base_dir):
learner.tmp_path = f'{base_dir}/tmp'
os.makedirs(learner.tmp_path, exist_ok=True)
learner.models_path = f'{base_dir}/models'
os.makedirs(learner.models_path, exist_ok=True)
# This is important for speed
cudnn.benchmark = True
global arg
args = parser.parse_args()
#print(args); exit()
if args.cycle_len > 1: args.cycle_len = int(args.cycle_len)
def main():
args.distributed = args.world_size > 1
args.gpu = 0
if args.distributed: args.gpu = args.rank % torch.cuda.device_count()
if args.distributed:
torch.cuda.set_device(args.gpu)
dist.init_process_group(backend=args.dist_backend, init_method=args.dist_url,
world_size=args.world_size)
if args.fp16: assert torch.backends.cudnn.enabled, "missing cudnn"
model = cifar10models.__dict__[args.arch] if args.arch in cifar10_names else models.__dict__[args.arch]
if args.pretrained: model = model(pretrained=True)
else: model = model()
model = model.cuda()
if args.distributed: model = DDP(model)
if args.data_parallel: model = nn.DataParallel(model, [0,1,2,3])
data, train_sampler = torch_loader(args.data, args.sz)
learner = Learner.from_model_data(model, data)
#print (learner.summary()); exit()
learner.crit = F.cross_entropy
learner.metrics = [accuracy]
if args.fp16: learner.half()
if args.prof: args.epochs,args.cycle_len = 1,0.01
if args.use_clr: args.use_clr = tuple(map(float, args.use_clr.split(',')))
# Full size
update_model_dir(learner, args.save_dir)
sargs = save_args('first_run', args.save_dir)
if args.warmup:
learner.fit(args.lr/10, 1, cycle_len=1, sampler=train_sampler, wds=args.weight_decay,
use_clr_beta=(100,1,0.9,0.8), loss_scale=args.loss_scale, **sargs)
learner.fit(args.lr,args.epochs, cycle_len=args.cycle_len,
sampler=train_sampler, wds=args.weight_decay,
use_clr_beta=args.use_clr, loss_scale=args.loss_scale,
**sargs)
save_sched(learner.sched, args.save_dir)
print('Finished!')
if args.use_tta:
log_preds,y = learner.TTA()
preds = np.mean(np.exp(log_preds),0)
acc = accuracy(torch.FloatTensor(preds),torch.LongTensor(y))
print('TTA acc:', acc)
with open(f'{args.save_dir}/tta_accuracy.txt', "a", 1) as f:
f.write(time.strftime("%Y-%m-%dT%H:%M:%S")+f"\tTTA accuracty: {acc}\n")
main()
