iicd/xautodl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

clean headers

Browse Source
This commit is contained in:
D-X-Y
2019-11-10 00:15:31 +11:00
parent 975fe4c385
commit fac556c176
41 changed files with 60 additions and 385 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
.gitignore vendored
View File

@@ -107,3 +107,6 @@ scripts-nas/.nfs00*
# logs and snapshots # logs and snapshots
output output
logs logs
# snapshot
a.pth

7
README.md
View File

@@ -5,7 +5,6 @@ This project contains the following neural architecture search algorithms, imple
- Network Pruning via Transformable Architecture Search, NeurIPS 2019 - Network Pruning via Transformable Architecture Search, NeurIPS 2019
- One-Shot Neural Architecture Search via Self-Evaluated Template Network, ICCV 2019 - One-Shot Neural Architecture Search via Self-Evaluated Template Network, ICCV 2019
- Searching for A Robust Neural Architecture in Four GPU Hours, CVPR 2019 - Searching for A Robust Neural Architecture in Four GPU Hours, CVPR 2019
- Auto-ReID: Searching for a Part-Aware ConvNet for Person Re-Identification, ICCV 2019
- several typical classification models, e.g., ResNet and DenseNet (see BASELINE.md) - several typical classification models, e.g., ResNet and DenseNet (see BASELINE.md)
@@ -104,12 +103,6 @@ CUDA_VISIBLE_DEVICES=0 bash ./scripts-search/algos/DARTS-V2.sh cifar10 -1
``` ```
## [Auto-ReID: Searching for a Part-Aware ConvNet for Person Re-Identification](https://arxiv.org/abs/1903.09776)
The part-aware module is defined at [here](https://github.com/D-X-Y/NAS-Projects/blob/master/lib/models/cell_searchs/operations.py#L85).
For more questions, please contact Ruijie Quan (Ruijie.Quan@student.uts.edu.au).
# Citation # Citation
If you find that this project helps your research, please consider citing some of the following papers: If you find that this project helps your research, please consider citing some of the following papers:

274
exps/AA-NAS-Bench-main.py
View File

@@ -1,274 +0,0 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, torch, random, argparse
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True
from copy import deepcopy
from pathlib import Path
lib_dir = (Path(__file__).parent / '..' / 'lib').resolve()
if str(lib_dir) not in sys.path: sys.path.insert(0, str(lib_dir))
from config_utils import load_config
from procedures import save_checkpoint, copy_checkpoint
from procedures import get_machine_info
from datasets import get_datasets
from log_utils import Logger, AverageMeter, time_string, convert_secs2time
from models import CellStructure, CellArchitectures, get_search_spaces
from AA_functions import evaluate_for_seed
def evaluate_all_datasets(arch, datasets, xpaths, splits, seed, arch_config, workers, logger):
machine_info, arch_config = get_machine_info(), deepcopy(arch_config)
all_infos = {'info': machine_info}
all_dataset_keys = []
# look all the datasets
for dataset, xpath, split in zip(datasets, xpaths, splits):
# train valid data
train_data, valid_data, xshape, class_num = get_datasets(dataset, xpath, -1)
# load the configurature
if dataset == 'cifar10' or dataset == 'cifar100':
config_path = 'configs/nas-benchmark/CIFAR.config'
split_info = load_config('configs/nas-benchmark/cifar-split.txt', None, None)
elif dataset.startswith('ImageNet16'):
config_path = 'configs/nas-benchmark/ImageNet-16.config'
split_info = load_config('configs/nas-benchmark/{:}-split.txt'.format(dataset), None, None)
else:
raise ValueError('invalid dataset : {:}'.format(dataset))
config = load_config(config_path, \
{'class_num': class_num,
'xshape' : xshape}, \
logger)
# check whether use splited validation set
if bool(split):
assert len(train_data) == len(split_info.train) + len(split_info.valid), 'invalid length : {:} vs {:} + {:}'.format(len(train_data), len(split_info.train), len(split_info.valid))
train_data_v2 = deepcopy(train_data)
train_data_v2.transform = valid_data.transform
valid_data = train_data_v2
# data loader
train_loader = torch.utils.data.DataLoader(train_data, batch_size=config.batch_size, sampler=torch.utils.data.sampler.SubsetRandomSampler(split_info.train), num_workers=workers, pin_memory=True)
valid_loader = torch.utils.data.DataLoader(valid_data, batch_size=config.batch_size, sampler=torch.utils.data.sampler.SubsetRandomSampler(split_info.valid), num_workers=workers, pin_memory=True)
else:
# data loader
train_loader = torch.utils.data.DataLoader(train_data, batch_size=config.batch_size, shuffle=True , num_workers=workers, pin_memory=True)
valid_loader = torch.utils.data.DataLoader(valid_data, batch_size=config.batch_size, shuffle=False, num_workers=workers, pin_memory=True)
dataset_key = '{:}'.format(dataset)
if bool(split): dataset_key = dataset_key + '-valid'
logger.log('Evaluate ||||||| {:10s} ||||||| Train-Num={:}, Valid-Num={:}, Train-Loader-Num={:}, Valid-Loader-Num={:}, batch size={:}'.format(dataset_key, len(train_data), len(valid_data), len(train_loader), len(valid_loader), config.batch_size))
logger.log('Evaluate ||||||| {:10s} ||||||| Config={:}'.format(dataset_key, config))
results = evaluate_for_seed(arch_config, config, arch, train_loader, valid_loader, seed, logger)
all_infos[dataset_key] = results
all_dataset_keys.append( dataset_key )
all_infos['all_dataset_keys'] = all_dataset_keys
return all_infos
def main(save_dir, workers, datasets, xpaths, splits, srange, arch_index, seeds, cover_mode, meta_info, arch_config):
assert torch.cuda.is_available(), 'CUDA is not available.'
torch.backends.cudnn.enabled = True
#torch.backends.cudnn.benchmark = True
torch.backends.cudnn.deterministic = True
torch.set_num_threads( workers )
assert len(srange) == 2 and 0 <= srange[0] <= srange[1], 'invalid srange : {:}'.format(srange)
sub_dir = Path(save_dir) / '{:06d}-{:06d}-C{:}-N{:}'.format(srange[0], srange[1], arch_config['channel'], arch_config['num_cells'])
logger = Logger(str(sub_dir), 0, False)
all_archs = meta_info['archs']
assert srange[1] < meta_info['total'], 'invalid range : {:}-{:} vs. {:}'.format(srange[0], srange[1], meta_info['total'])
assert arch_index == -1 or srange[0] <= arch_index <= srange[1], 'invalid range : {:} vs. {:} vs. {:}'.format(srange[0], arch_index, srange[1])
if arch_index == -1:
to_evaluate_indexes = list(range(srange[0], srange[1]+1))
else:
to_evaluate_indexes = [arch_index]
logger.log('xargs : seeds = {:}'.format(seeds))
logger.log('xargs : arch_index = {:}'.format(arch_index))
logger.log('xargs : cover_mode = {:}'.format(cover_mode))
logger.log('-'*100)
logger.log('Start evaluating range =: {:06d} vs. {:06d} vs. {:06d} / {:06d} with cover-mode={:}'.format(srange[0], arch_index, srange[1], meta_info['total'], cover_mode))
for i, (dataset, xpath, split) in enumerate(zip(datasets, xpaths, splits)):
logger.log('--->>> Evaluate {:}/{:} : dataset={:9s}, path={:}, split={:}'.format(i, len(datasets), dataset, xpath, split))
logger.log('--->>> architecture config : {:}'.format(arch_config))
start_time, epoch_time = time.time(), AverageMeter()
for i, index in enumerate(to_evaluate_indexes):
arch = all_archs[index]
logger.log('\n{:} evaluate {:06d}/{:06d} ({:06d}/{:06d})-th architecture [seeds={:}] {:}'.format('-'*15, i, len(to_evaluate_indexes), index, meta_info['total'], seeds, '-'*15))
#logger.log('{:} {:} {:}'.format('-'*15, arch.tostr(), '-'*15))
logger.log('{:} {:} {:}'.format('-'*15, arch, '-'*15))
# test this arch on different datasets with different seeds
has_continue = False
for seed in seeds:
to_save_name = sub_dir / 'arch-{:06d}-seed-{:04d}.pth'.format(index, seed)
if to_save_name.exists():
if cover_mode:
logger.log('Find existing file : {:}, remove it before evaluation'.format(to_save_name))
os.remove(str(to_save_name))
else :
logger.log('Find existing file : {:}, skip this evaluation'.format(to_save_name))
has_continue = True
continue
results = evaluate_all_datasets(CellStructure.str2structure(arch), \
datasets, xpaths, splits, seed, \
arch_config, workers, logger)
torch.save(results, to_save_name)
logger.log('{:} valuate {:06d}/{:06d} ({:06d}/{:06d})-th seed={:} done, save into {:}'.format('-'*15, i, len(to_evaluate_indexes), index, meta_info['total'], seed, to_save_name))
# measure elapsed time
if not has_continue: epoch_time.update(time.time() - start_time)
start_time = time.time()
need_time = 'Time Left: {:}'.format( convert_secs2time(epoch_time.avg * (len(to_evaluate_indexes)-i-1), True) )
logger.log('This arch costs : {:}'.format( convert_secs2time(epoch_time.val, True) ))
logger.log('{:}'.format('*'*100))
logger.log('{:} {:74s} {:}'.format('*'*10, '{:06d}/{:06d} ({:06d}/{:06d})-th done, left {:}'.format(i, len(to_evaluate_indexes), index, meta_info['total'], need_time), '*'*10))
logger.log('{:}'.format('*'*100))
logger.close()
def train_single_model(save_dir, workers, datasets, xpaths, splits, seeds, model_str, arch_config):
assert torch.cuda.is_available(), 'CUDA is not available.'
torch.backends.cudnn.enabled = True
torch.backends.cudnn.deterministic = True
#torch.backends.cudnn.benchmark = True
torch.set_num_threads( workers )
save_dir = Path(save_dir) / 'specifics' / '{:}-{:}-{:}'.format(model_str, arch_config['channel'], arch_config['num_cells'])
logger = Logger(str(save_dir), 0, False)
if model_str in CellArchitectures:
arch = CellArchitectures[model_str]
logger.log('The model string is found in pre-defined architecture dict : {:}'.format(model_str))
else:
try:
arch = CellStructure.str2structure(model_str)
except:
raise ValueError('Invalid model string : {:}. It can not be found or parsed.'.format(model_str))
assert arch.check_valid_op(get_search_spaces('cell', 'full')), '{:} has the invalid op.'.format(arch)
logger.log('Start train-evaluate {:}'.format(arch.tostr()))
logger.log('arch_config : {:}'.format(arch_config))
start_time, seed_time = time.time(), AverageMeter()
for _is, seed in enumerate(seeds):
logger.log('\nThe {:02d}/{:02d}-th seed is {:} ----------------------<.>----------------------'.format(_is, len(seeds), seed))
to_save_name = save_dir / 'seed-{:04d}.pth'.format(seed)
if to_save_name.exists():
logger.log('Find the existing file {:}, directly load!'.format(to_save_name))
checkpoint = torch.load(to_save_name)
else:
logger.log('Does not find the existing file {:}, train and evaluate!'.format(to_save_name))
checkpoint = evaluate_all_datasets(arch, datasets, xpaths, splits, seed, arch_config, workers, logger)
torch.save(checkpoint, to_save_name)
# log information
logger.log('{:}'.format(checkpoint['info']))
all_dataset_keys = checkpoint['all_dataset_keys']
for dataset_key in all_dataset_keys:
logger.log('\n{:} dataset : {:} {:}'.format('-'*15, dataset_key, '-'*15))
dataset_info = checkpoint[dataset_key]
#logger.log('Network ==>\n{:}'.format( dataset_info['net_string'] ))
logger.log('Flops = {:} MB, Params = {:} MB'.format(dataset_info['flop'], dataset_info['param']))
logger.log('config : {:}'.format(dataset_info['config']))
logger.log('Training State (finish) = {:}'.format(dataset_info['finish-train']))
last_epoch = dataset_info['total_epoch'] - 1
train_acc1es, train_acc5es = dataset_info['train_acc1es'], dataset_info['train_acc5es']
valid_acc1es, valid_acc5es = dataset_info['valid_acc1es'], dataset_info['valid_acc5es']
logger.log('Last Info : Train = Acc@1 {:.2f}% Acc@5 {:.2f}% Error@1 {:.2f}%, Test = Acc@1 {:.2f}% Acc@5 {:.2f}% Error@1 {:.2f}%'.format(train_acc1es[last_epoch], train_acc5es[last_epoch], 100-train_acc1es[last_epoch], valid_acc1es[last_epoch], valid_acc5es[last_epoch], 100-valid_acc1es[last_epoch]))
# measure elapsed time
seed_time.update(time.time() - start_time)
start_time = time.time()
need_time = 'Time Left: {:}'.format( convert_secs2time(seed_time.avg * (len(seeds)-_is-1), True) )
logger.log('\n<<<***>>> The {:02d}/{:02d}-th seed is {:} <finish> other procedures need {:}'.format(_is, len(seeds), seed, need_time))
logger.close()
def generate_meta_info(save_dir, max_node, divide=40):
aa_nas_bench_ss = get_search_spaces('cell', 'aa-nas')
archs = CellStructure.gen_all(aa_nas_bench_ss, max_node, False)
print ('There are {:} archs vs {:}.'.format(len(archs), len(aa_nas_bench_ss) ** ((max_node-1)*max_node/2)))
random.seed( 88 ) # please do not change this line for reproducibility
random.shuffle( archs )
# to test fixed-random shuffle
#print ('arch [0] : {:}\n---->>>> {:}'.format( archs[0], archs[0].tostr() ))
#print ('arch [9] : {:}\n---->>>> {:}'.format( archs[9], archs[9].tostr() ))
assert archs[0 ].tostr() == '|avg_pool_3x3~0|+|nor_conv_1x1~0|skip_connect~1|+|nor_conv_1x1~0|skip_connect~1|skip_connect~2|', 'please check the 0-th architecture : {:}'.format(archs[0])
assert archs[9 ].tostr() == '|avg_pool_3x3~0|+|none~0|none~1|+|skip_connect~0|none~1|nor_conv_3x3~2|', 'please check the 9-th architecture : {:}'.format(archs[9])
assert archs[123].tostr() == '|avg_pool_3x3~0|+|avg_pool_3x3~0|nor_conv_1x1~1|+|none~0|avg_pool_3x3~1|nor_conv_3x3~2|', 'please check the 123-th architecture : {:}'.format(archs[123])
total_arch = len(archs)
num = 50000
indexes_5W = list(range(num))
random.seed( 1021 )
random.shuffle( indexes_5W )
train_split = sorted( list(set(indexes_5W[:num//2])) )
valid_split = sorted( list(set(indexes_5W[num//2:])) )
assert len(train_split) + len(valid_split) == num
assert train_split[0] == 0 and train_split[10] == 26 and train_split[111] == 203 and valid_split[0] == 1 and valid_split[10] == 18 and valid_split[111] == 242, '{:} {:} {:} - {:} {:} {:}'.format(train_split[0], train_split[10], train_split[111], valid_split[0], valid_split[10], valid_split[111])
splits = {num: {'train': train_split, 'valid': valid_split} }
info = {'archs' : [x.tostr() for x in archs],
'total' : total_arch,
'max_node' : max_node,
'splits': splits}
save_dir = Path(save_dir)
save_dir.mkdir(parents=True, exist_ok=True)
save_name = save_dir / 'meta-node-{:}.pth'.format(max_node)
assert not save_name.exists(), '{:} already exist'.format(save_name)
torch.save(info, save_name)
print ('save the meta file into {:}'.format(save_name))
script_name = save_dir / 'meta-node-{:}.script.txt'.format(max_node)
with open(str(script_name), 'w') as cfile:
gaps = total_arch // divide
for start in range(0, total_arch, gaps):
xend = min(start+gaps, total_arch)
cfile.write('bash ./scripts-search/AA-NAS-train-archs.sh {:5d} {:5d} -1 \'777 888 999\'\n'.format(start, xend-1))
print ('save the training script into {:}'.format(script_name))
if __name__ == '__main__':
#mode_choices = ['meta', 'new', 'cover'] + ['specific-{:}'.format(_) for _ in CellArchitectures.keys()]
parser = argparse.ArgumentParser(description='Algorithm-Agnostic NAS Benchmark', formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--mode' , type=str, required=True, help='The script mode.')
parser.add_argument('--save_dir', type=str, help='Folder to save checkpoints and log.')
parser.add_argument('--max_node', type=int, help='The maximum node in a cell.')
# use for train the model
parser.add_argument('--workers', type=int, default=8, help='number of data loading workers (default: 2)')
parser.add_argument('--srange' , type=int, nargs='+', help='The range of models to be evaluated')
parser.add_argument('--arch_index', type=int, default=-1, help='The architecture index to be evaluated (cover mode).')
parser.add_argument('--datasets', type=str, nargs='+', help='The applied datasets.')
parser.add_argument('--xpaths', type=str, nargs='+', help='The root path for this dataset.')
parser.add_argument('--splits', type=int, nargs='+', help='The root path for this dataset.')
parser.add_argument('--seeds' , type=int, nargs='+', help='The range of models to be evaluated')
parser.add_argument('--channel', type=int, help='The number of channels.')
parser.add_argument('--num_cells', type=int, help='The number of cells in one stage.')
args = parser.parse_args()
assert args.mode in ['meta', 'new', 'cover'] or args.mode.startswith('specific-'), 'invalid mode : {:}'.format(args.mode)
if args.mode == 'meta':
generate_meta_info(args.save_dir, args.max_node)
elif args.mode.startswith('specific'):
assert len(args.mode.split('-')) == 2, 'invalid mode : {:}'.format(args.mode)
model_str = args.mode.split('-')[1]
train_single_model(args.save_dir, args.workers, args.datasets, args.xpaths, args.splits, \
tuple(args.seeds), model_str, {'channel': args.channel, 'num_cells': args.num_cells})
else:
meta_path = Path(args.save_dir) / 'meta-node-{:}.pth'.format(args.max_node)
assert meta_path.exists(), '{:} does not exist.'.format(meta_path)
meta_info = torch.load( meta_path )
# check whether args is ok
assert len(args.srange) == 2 and args.srange[0] <= args.srange[1], 'invalid length of srange args: {:}'.format(args.srange)
assert len(args.seeds) > 0, 'invalid length of seeds args: {:}'.format(args.seeds)
assert len(args.datasets) == len(args.xpaths) == len(args.splits), 'invalid infos : {:} vs {:} vs {:}'.format(len(args.datasets), len(args.xpaths), len(args.splits))
assert args.workers > 0, 'invalid number of workers : {:}'.format(args.workers)
main(args.save_dir, args.workers, args.datasets, args.xpaths, args.splits, \
tuple(args.srange), args.arch_index, tuple(args.seeds), \
args.mode == 'cover', meta_info, \
{'channel': args.channel, 'num_cells': args.num_cells})

38
exps/AA-NAS-test-API.py Normal file
View File

@@ -0,0 +1,38 @@
import os, sys, time, queue, torch
from pathlib import Path
lib_dir = (Path(__file__).parent / '..' / 'lib').resolve()
if str(lib_dir) not in sys.path: sys.path.insert(0, str(lib_dir))
from log_utils import time_string
from models import CellStructure
def get_unique_matrix(archs, consider_zero):
UniquStrs = [arch.to_unique_str(consider_zero) for arch in archs]
print ('{:} create unique-string done'.format(time_string()))
sm_matrix = torch.eye(len(archs)).bool()
for i, _ in enumerate(UniquStrs):
for j in range(i):
sm_matrix[i,j] = sm_matrix[j,i] = UniquStrs[i] == UniquStrs[j]
unique_ids, unique_num = [-1 for _ in archs], 0
for i in range(len(unique_ids)):
if unique_ids[i] > -1: continue
neighbours = sm_matrix[i].nonzero().view(-1).tolist()
for nghb in neighbours:
assert unique_ids[nghb] == -1, 'impossible'
unique_ids[nghb] = unique_num
unique_num += 1
return sm_matrix, unique_ids, unique_num
def check_unique_arch():
print ('{:} start'.format(time_string()))
meta_info = torch.load('./output/AA-NAS-BENCH-4/meta-node-4.pth')
arch_strs = meta_info['archs']
archs = [CellStructure.str2structure(arch_str) for arch_str in arch_strs]
_, _, unique_num = get_unique_matrix(archs, False)
print ('{:} There are {:} unique architectures (not considering zero).'.format(time_string(), unique_num))
_, _, unique_num = get_unique_matrix(archs, True)
print ('{:} There are {:} unique architectures (considering zero).'.format(time_string(), unique_num))
if __name__ == '__main__':
check_unique_arch()

3
lib/config_utils/__init__.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
from .configure_utils import load_config, dict2config, configure2str from .configure_utils import load_config, dict2config, configure2str
from .basic_args import obtain_basic_args from .basic_args import obtain_basic_args
from .attention_args import obtain_attention_args from .attention_args import obtain_attention_args

3
lib/config_utils/attention_args.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
from .share_args import add_shared_args from .share_args import add_shared_args

3
lib/config_utils/basic_args.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
from .share_args import add_shared_args from .share_args import add_shared_args

3
lib/config_utils/cls_init_args.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
from .share_args import add_shared_args from .share_args import add_shared_args

3
lib/config_utils/cls_kd_args.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
from .share_args import add_shared_args from .share_args import add_shared_args

3
lib/config_utils/pruning_args.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
from .share_args import add_shared_args from .share_args import add_shared_args

3
lib/config_utils/random_baseline.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
from .share_args import add_shared_args from .share_args import add_shared_args

3
lib/config_utils/search_args.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
from .share_args import add_shared_args from .share_args import add_shared_args

3
lib/config_utils/search_single_args.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
from .share_args import add_shared_args from .share_args import add_shared_args

3
lib/config_utils/share_args.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, random, argparse import os, sys, time, random, argparse
def add_shared_args( parser ): def add_shared_args( parser ):

3
lib/datasets/__init__.py
View File

@@ -1,5 +1,2 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
from .get_dataset_with_transform import get_datasets from .get_dataset_with_transform import get_datasets
from .SearchDatasetWrap import SearchDataset from .SearchDatasetWrap import SearchDataset

3
lib/datasets/get_dataset_with_transform.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, torch import os, sys, torch
import os.path as osp import os.path as osp
import numpy as np import numpy as np

3
lib/log_utils/__init__.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
from .logger import Logger from .logger import Logger
from .print_logger import PrintLogger from .print_logger import PrintLogger
from .meter import AverageMeter from .meter import AverageMeter

3
lib/log_utils/meter.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import time, sys import time, sys
import numpy as np import numpy as np

6
lib/log_utils/print_logger.py
View File

@@ -1,8 +1,4 @@
################################################## import os, sys, time
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import importlib, warnings
import os, sys, time, numpy as np
class PrintLogger(object): class PrintLogger(object):

3
lib/models/CifarResNet.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import torch import torch
import torch.nn as nn import torch.nn as nn
import torch.nn.functional as F import torch.nn.functional as F

3
lib/models/__init__.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import torch import torch
from os import path as osp from os import path as osp

3
lib/models/cell_infers/cells.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import torch import torch
import torch.nn as nn import torch.nn as nn
from copy import deepcopy from copy import deepcopy

3
lib/models/cell_infers/tiny_network.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import torch import torch
import torch.nn as nn import torch.nn as nn
from ..cell_operations import ResNetBasicblock from ..cell_operations import ResNetBasicblock

18
lib/models/cell_searchs/genotypes.py
View File

@@ -60,6 +60,24 @@ class Structure:
strings.append( string ) strings.append( string )
return '+'.join(strings) return '+'.join(strings)
def to_unique_str(self, consider_zero=False):
# this is used to identify the isomorphic cell, which rerquires the prior knowledge of operation
# two operations are special, i.e., none and skip_connect
nodes = {0: '0'}
for i_node, node_info in enumerate(self.nodes):
cur_node = []
for op, xin in node_info:
if consider_zero:
if op == 'none' or nodes[xin] == '#': x = '#' # zero
elif op == 'skip_connect': x = nodes[xin]
else: x = nodes[xin] + '@{:}'.format(op)
else:
if op == 'skip_connect': x = nodes[xin]
else: x = nodes[xin] + '@{:}'.format(op)
cur_node.append(x)
nodes[i_node+1] = '+'.join( sorted(cur_node) )
return nodes[ len(self.nodes) ]
def check_valid_op(self, op_names): def check_valid_op(self, op_names):
for node_info in self.nodes: for node_info in self.nodes:
for inode_edge in node_info: for inode_edge in node_info:

2
lib/models/cell_searchs/search_model_darts_v1.py
View File

@@ -1,5 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
######################################################## ########################################################
# DARTS: Differentiable Architecture Search, ICLR 2019 # # DARTS: Differentiable Architecture Search, ICLR 2019 #
######################################################## ########################################################

2
lib/models/cell_searchs/search_model_darts_v2.py
View File

@@ -1,5 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
######################################################## ########################################################
# DARTS: Differentiable Architecture Search, ICLR 2019 # # DARTS: Differentiable Architecture Search, ICLR 2019 #
######################################################## ########################################################

2
lib/models/cell_searchs/search_model_enas.py
View File

@@ -1,5 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
########################################################################## ##########################################################################
# Efficient Neural Architecture Search via Parameters Sharing, ICML 2018 # # Efficient Neural Architecture Search via Parameters Sharing, ICML 2018 #
########################################################################## ##########################################################################

2
lib/models/cell_searchs/search_model_enas_utils.py
View File

@@ -1,5 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
########################################################################## ##########################################################################
# Efficient Neural Architecture Search via Parameters Sharing, ICML 2018 # # Efficient Neural Architecture Search via Parameters Sharing, ICML 2018 #
########################################################################## ##########################################################################

2
lib/models/cell_searchs/search_model_gdas.py
View File

@@ -1,5 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
########################################################################### ###########################################################################
# Searching for A Robust Neural Architecture in Four GPU Hours, CVPR 2019 # # Searching for A Robust Neural Architecture in Four GPU Hours, CVPR 2019 #
########################################################################### ###########################################################################

2
lib/models/cell_searchs/search_model_setn.py
View File

@@ -1,5 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
###################################################################################### ######################################################################################
# One-Shot Neural Architecture Search via Self-Evaluated Template Network, ICCV 2019 # # One-Shot Neural Architecture Search via Self-Evaluated Template Network, ICCV 2019 #
###################################################################################### ######################################################################################

2
lib/models/shape_infers/shared_utils.py
View File

@@ -1,5 +1,3 @@
# Xuanyi Dong
def parse_channel_info(xstring): def parse_channel_info(xstring):
blocks = xstring.split(' ') blocks = xstring.split(' ')
blocks = [x.split('-') for x in blocks] blocks = [x.split('-') for x in blocks]

3
lib/nas_infer_model/DXYs/genotypes.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
from collections import namedtuple from collections import namedtuple
Genotype = namedtuple('Genotype', 'normal normal_concat reduce reduce_concat connectN connects') Genotype = namedtuple('Genotype', 'normal normal_concat reduce reduce_concat connectN connects')

3
lib/nas_infer_model/__init__.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import torch import torch
def obtain_nas_infer_model(config): def obtain_nas_infer_model(config):

3
lib/procedures/__init__.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
from .starts import prepare_seed, prepare_logger, get_machine_info, save_checkpoint, copy_checkpoint from .starts import prepare_seed, prepare_logger, get_machine_info, save_checkpoint, copy_checkpoint
from .optimizers import get_optim_scheduler from .optimizers import get_optim_scheduler

3
lib/procedures/basic_main.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, torch import os, sys, time, torch
from log_utils import AverageMeter, time_string from log_utils import AverageMeter, time_string
from utils import obtain_accuracy from utils import obtain_accuracy

3
lib/procedures/optimizers.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import math, torch import math, torch
import torch.nn as nn import torch.nn as nn
from bisect import bisect_right from bisect import bisect_right

3
lib/procedures/search_main.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, torch import os, sys, time, torch
from log_utils import AverageMeter, time_string from log_utils import AverageMeter, time_string
from utils import obtain_accuracy from utils import obtain_accuracy

3
lib/procedures/search_main_v2.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, torch import os, sys, time, torch
from log_utils import AverageMeter, time_string from log_utils import AverageMeter, time_string
from utils import obtain_accuracy from utils import obtain_accuracy

3
lib/procedures/simple_KD_main.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
import os, sys, time, torch import os, sys, time, torch
import torch.nn.functional as F import torch.nn.functional as F
# our modules # our modules

3
lib/utils/__init__.py
View File

@@ -1,6 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
from .evaluation_utils import obtain_accuracy from .evaluation_utils import obtain_accuracy
from .gpu_manager import GPUManager from .gpu_manager import GPUManager
from .flop_benchmark import get_model_infos from .flop_benchmark import get_model_infos

4
lib/utils/flop_benchmark.py
View File

@@ -1,7 +1,3 @@
##################################################
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2019 #
##################################################
# modified from https://github.com/warmspringwinds/pytorch-segmentation-detection/blob/master/pytorch_segmentation_detection/utils/flops_benchmark.py
import copy, torch import copy, torch
import torch.nn as nn import torch.nn as nn
import numpy as np import numpy as np