def use(args):
'''\
Transform images with the net.
Args:
args: namespace of arguments. Run 'artRecycle use --help' for info.
'''
# Model name and paths
model_name = '{}|{}'.format(*args.datasets)
model_path, log_path, logs_path = _prepare_directories(model_name,
resume=True)
model_json = os.path.join(model_path, 'keras.json')
model_checkpoint = os.path.join(model_path, 'model')
# Define dataset
image_shape = (300, 300, 3)
test_dataset, test_size = data.load_pair(*args.datasets,
'test',
shape=image_shape,
batch=args.batch)
# Define keras model
keras_model, model_layer = models.define_model(image_shape)
# Load
keras_model.load_weights(model_checkpoint)
print('> Weights loaded')
raise NotImplementedError()
def test(opt):
# logging
logger = base_utils.get_logger('base')
if opt['verbose']:
logger.info('{} Configurations {}'.format('=' * 20, '=' * 20))
base_utils.print_options(opt, logger)
# infer and evaluate performance for each model
for load_path in opt['model']['generator']['load_path_lst']:
# setup model index
model_idx = osp.splitext(osp.split(load_path)[-1])[0]
# log
logger.info('=' * 40)
logger.info('Testing model: {}'.format(model_idx))
logger.info('=' * 40)
# create model
opt['model']['generator']['load_path'] = load_path
model = define_model(opt)
model_idx = osp.basename(opt['model']['generator']['load_path']).split('.')[0]
# for each test dataset
for dataset_idx in sorted(opt['dataset'].keys()):
# use dataset with prefix `test`
if not dataset_idx.startswith('test'):
continue
validate(opt, model, logger, dataset_idx, model_idx, compute_metrics=False)
logger.info('-' * 40)
# logging
logger.info('Finish testing')
logger.info('=' * 40)
def test(opt):
# logging
logger = base_utils.get_logger('base')
if opt['verbose']:
logger.info('{} Configurations {}'.format('=' * 20, '=' * 20))
base_utils.print_options(opt, logger)
# infer and evaluate performance for each model
for load_path in opt['model']['generator']['load_path_lst']:
# setup model index
model_idx = osp.splitext(osp.split(load_path)[-1])[0]
# log
logger.info('=' * 40)
logger.info('Testing model: {}'.format(model_idx))
logger.info('=' * 40)
# create model
opt['model']['generator']['load_path'] = load_path
model = define_model(opt)
# for each test dataset
for dataset_idx in sorted(opt['dataset'].keys()):
# use dataset with prefix `test`
if not dataset_idx.startswith('test'):
continue
ds_name = opt['dataset'][dataset_idx]['name']
logger.info('Testing on {}: {}'.format(dataset_idx, ds_name))
# create data loader
test_loader = create_dataloader(opt, dataset_idx=dataset_idx)
# infer and store results for each sequence
for i, data in enumerate(test_loader):
# fetch data
lr_data = data['lr'][0]
seq_idx = data['seq_idx'][0]
frm_idx = [frm_idx[0] for frm_idx in data['frm_idx']]
# infer
hr_seq = model.infer(lr_data) # thwc|rgb|uint8
# save results (optional)
if opt['test']['save_res']:
res_dir = osp.join(opt['test']['res_dir'], ds_name,
model_idx)
res_seq_dir = osp.join(res_dir, seq_idx)
data_utils.save_sequence(res_seq_dir,
hr_seq,
frm_idx,
to_bgr=True)
logger.info('-' * 40)
# logging
logger.info('Finish testing')
logger.info('=' * 40)
def debug(args):
'''\
Debugging function.
Args:
args: namespace of arguments. Run --help for info.
'''
import matplotlib.pyplot as plt
print('> Debug')
# Saving the Tensorboard graph without training
# Model
image_shape = (300, 300, 3)
keras_model, model_layer = models.define_model(image_shape)
keras_model.summary()
# TensorBoard callback writer
tbCallback = tf.keras.callbacks.TensorBoard('debug', write_graph=True)
tbCallback.set_model(keras_model)
def train(opt):
# logging
logger = base_utils.get_logger('base')
logger.info('{} Options {}'.format('='*20, '='*20))
base_utils.print_options(opt, logger)
# create data loader
train_loader = create_dataloader(opt, dataset_idx='train')
# create downsampling kernels for BD degradation
kernel = data_utils.create_kernel(opt)
# create model
model = define_model(opt)
# training configs
total_sample = len(train_loader.dataset)
iter_per_epoch = len(train_loader)
total_iter = opt['train']['total_iter']
total_epoch = int(math.ceil(total_iter / iter_per_epoch))
start_iter, iter = opt['train']['start_iter'], 0
test_freq = opt['test']['test_freq']
log_freq = opt['logger']['log_freq']
ckpt_freq = opt['logger']['ckpt_freq']
logger.info('Number of training samples: {}'.format(total_sample))
logger.info('Total epochs needed: {} for {} iterations'.format(
total_epoch, total_iter))
# train
for epoch in range(total_epoch):
for data in train_loader:
# update iter
iter += 1
curr_iter = start_iter + iter
if iter > total_iter:
logger.info('Finish training')
break
# update learning rate
model.update_learning_rate()
# prepare data
data = prepare_data(opt, data, kernel)
# train for a mini-batch
model.train(data)
# update running log
model.update_running_log()
# log
if log_freq > 0 and iter % log_freq == 0:
# basic info
msg = '[epoch: {} | iter: {}'.format(epoch, curr_iter)
for lr_type, lr in model.get_current_learning_rate().items():
msg += ' | {}: {:.2e}'.format(lr_type, lr)
msg += '] '
# loss info
log_dict = model.get_running_log()
msg += ', '.join([
'{}: {:.3e}'.format(k, v) for k, v in log_dict.items()])
logger.info(msg)
# save model
if ckpt_freq > 0 and iter % ckpt_freq == 0:
model.save(curr_iter)
# evaluate performance
if test_freq > 0 and iter % test_freq == 0:
# setup model index
model_idx = 'G_iter{}'.format(curr_iter)
# for each testset
for dataset_idx in sorted(opt['dataset'].keys()):
# use dataset with prefix `test`
if not dataset_idx.startswith('test'):
continue
ds_name = opt['dataset'][dataset_idx]['name']
logger.info(
'Testing on {}: {}'.format(dataset_idx, ds_name))
# create data loader
test_loader = create_dataloader(opt, dataset_idx=dataset_idx)
# define metric calculator
metric_calculator = MetricCalculator(opt)
# infer and compute metrics for each sequence
for data in test_loader:
# fetch data
lr_data = data['lr'][0]
seq_idx = data['seq_idx'][0]
frm_idx = [frm_idx[0] for frm_idx in data['frm_idx']]
# infer
hr_seq = model.infer(lr_data) # thwc|rgb|uint8
# save results (optional)
if opt['test']['save_res']:
res_dir = osp.join(
opt['test']['res_dir'], ds_name, model_idx)
res_seq_dir = osp.join(res_dir, seq_idx)
data_utils.save_sequence(
res_seq_dir, hr_seq, frm_idx, to_bgr=True)
# compute metrics for the current sequence
true_seq_dir = osp.join(
opt['dataset'][dataset_idx]['gt_seq_dir'], seq_idx)
metric_calculator.compute_sequence_metrics(
seq_idx, true_seq_dir, '', pred_seq=hr_seq)
# save/print metrics
if opt['test'].get('save_json'):
# save results to json file
json_path = osp.join(
opt['test']['json_dir'], '{}_avg.json'.format(ds_name))
metric_calculator.save_results(
model_idx, json_path, override=True)
else:
# print directly
metric_calculator.display_results()
def test(opt):
# logging
logger = base_utils.get_logger('base')
if opt['verbose']:
logger.info('{} Configurations {}'.format('=' * 20, '=' * 20))
base_utils.print_options(opt, logger)
# infer and evaluate performance for each model
for load_path in opt['model']['generator']['load_path_lst']:
# setup model index
model_idx = osp.splitext(osp.split(load_path)[-1])[0]
# log
logger.info('=' * 40)
logger.info('Testing model: {}'.format(model_idx))
logger.info('=' * 40)
# create model
opt['model']['generator']['load_path'] = load_path
model = define_model(opt)
# for each test dataset
for dataset_idx in sorted(opt['dataset'].keys()):
# use dataset with prefix `test`
if not dataset_idx.startswith('test'):
continue
ds_name = opt['dataset'][dataset_idx]['name']
logger.info('Testing on {}: {}'.format(dataset_idx, ds_name))
# define metric calculator
try:
metric_calculator = MetricCalculator(opt)
except:
print('No metirc need to compute!')
# create data loader
test_loader = create_dataloader(opt, dataset_idx=dataset_idx)
# infer and store results for each sequence
for i, data in enumerate(test_loader):
# fetch data
lr_data = data['lr'][0]
seq_idx = data['seq_idx'][0]
frm_idx = [frm_idx[0] for frm_idx in data['frm_idx']]
# infer
hr_seq = model.infer(lr_data) # thwc|rgb|uint8
# save results (optional)
if opt['test']['save_res']:
res_dir = osp.join(opt['test']['res_dir'], ds_name,
model_idx)
res_seq_dir = osp.join(res_dir, seq_idx)
data_utils.save_sequence(res_seq_dir,
hr_seq,
frm_idx,
to_bgr=True)
# compute metrics for the current sequence
true_seq_dir = osp.join(
opt['dataset'][dataset_idx]['gt_seq_dir'], seq_idx)
try:
metric_calculator.compute_sequence_metrics(seq_idx,
true_seq_dir,
'',
pred_seq=hr_seq)
except:
print('No metirc need to compute!')
# save/print metrics
try:
if opt['test'].get('save_json'):
# save results to json file
json_path = osp.join(opt['test']['json_dir'],
'{}_avg.json'.format(ds_name))
metric_calculator.save_results(model_idx,
json_path,
override=True)
else:
# print directly
metric_calculator.display_results()
except:
print('No metirc need to save!')
logger.info('-' * 40)
# logging
logger.info('Finish testing')
logger.info('=' * 40)
def test(opt):
# logging
base_utils.print_options(opt)
# infer and evaluate performance for each model
for load_path in opt['model']['generator']['load_path_lst']:
# set model index
model_idx = osp.splitext(osp.split(load_path)[-1])[0]
# log
base_utils.log_info(f'{"=" * 40}')
base_utils.log_info(f'Testing model: {model_idx}')
base_utils.log_info(f'{"=" * 40}')
# create model
opt['model']['generator']['load_path'] = load_path
model = define_model(opt)
# for each test dataset
for dataset_idx in sorted(opt['dataset'].keys()):
# select testing dataset
if 'test' not in dataset_idx:
continue
ds_name = opt['dataset'][dataset_idx]['name']
base_utils.log_info(f'Testing on {ds_name} dataset')
# create data loader
test_loader = create_dataloader(opt, phase='test', idx=dataset_idx)
test_dataset = test_loader.dataset
num_seq = len(test_dataset)
# create metric calculator
metric_calculator = create_metric_calculator(opt)
# infer a sequence
rank, world_size = dist_utils.get_dist_info()
for idx in range(rank, num_seq, world_size):
# fetch data
data = test_dataset[idx]
# prepare data
model.prepare_inference_data(data)
# infer
hr_seq = model.infer()
# save hr results
if opt['test']['save_res']:
res_dir = osp.join(opt['test']['res_dir'], ds_name,
model_idx)
res_seq_dir = osp.join(res_dir, data['seq_idx'])
data_utils.save_sequence(res_seq_dir,
hr_seq,
data['frm_idx'],
to_bgr=True)
# compute metrics for the current sequence
if metric_calculator is not None:
gt_seq = data['gt'].numpy()
metric_calculator.compute_sequence_metrics(
data['seq_idx'], gt_seq, hr_seq)
# save/print results
if metric_calculator is not None:
seq_idx_lst = [data['seq_idx'] for data in test_dataset]
metric_calculator.gather(seq_idx_lst)
if opt['test'].get('save_json'):
# write results to a json file
json_path = osp.join(opt['test']['json_dir'],
f'{ds_name}_avg.json')
metric_calculator.save(model_idx, json_path, override=True)
else:
# print directly
metric_calculator.display()
base_utils.log_info('-' * 40)
def train(opt):
# print configurations
base_utils.log_info(f'{20*"-"} Configurations {20*"-"}')
base_utils.print_options(opt)
# create data loader
train_loader = create_dataloader(opt, phase='train', idx='train')
# build model
model = define_model(opt)
# set training params
total_sample, iter_per_epoch = len(train_loader.dataset), len(train_loader)
total_iter = opt['train']['total_iter']
total_epoch = int(math.ceil(total_iter / iter_per_epoch))
start_iter, iter = opt['train']['start_iter'], 0
test_freq = opt['test']['test_freq']
log_freq = opt['logger']['log_freq']
ckpt_freq = opt['logger']['ckpt_freq']
base_utils.log_info(f'Number of the training samples: {total_sample}')
base_utils.log_info(
f'{total_epoch} epochs needed for {total_iter} iterations')
# train
for epoch in range(total_epoch):
if opt['dist']:
train_loader.sampler.set_epoch(epoch)
for data in train_loader:
# update iter
iter += 1
curr_iter = start_iter + iter
if iter > total_iter: break
# prepare data
model.prepare_training_data(data)
# train a mini-batch
model.train()
# update running log
model.update_running_log()
# update learning rate
model.update_learning_rate()
# print messages
if log_freq > 0 and curr_iter % log_freq == 0:
msg = model.get_format_msg(epoch, curr_iter)
base_utils.log_info(msg)
# save model
if ckpt_freq > 0 and curr_iter % ckpt_freq == 0:
model.save(curr_iter)
# evaluate model
if test_freq > 0 and curr_iter % test_freq == 0:
# set model index
model_idx = f'G_iter{curr_iter}'
# for each testset
for dataset_idx in sorted(opt['dataset'].keys()):
# select test dataset
if 'test' not in dataset_idx: continue
ds_name = opt['dataset'][dataset_idx]['name']
base_utils.log_info(f'Testing on {ds_name} dataset')
# create data loader
test_loader = create_dataloader(opt,
phase='test',
idx=dataset_idx)
test_dataset = test_loader.dataset
num_seq = len(test_dataset)
# create metric calculator
metric_calculator = create_metric_calculator(opt)
# infer a sequence
rank, world_size = dist_utils.get_dist_info()
for idx in range(rank, num_seq, world_size):
# fetch data
data = test_dataset[idx]
# prepare data
model.prepare_inference_data(data)
# infer
hr_seq = model.infer()
# save hr results
if opt['test']['save_res']:
res_dir = osp.join(opt['test']['res_dir'], ds_name,
model_idx)
res_seq_dir = osp.join(res_dir, data['seq_idx'])
data_utils.save_sequence(res_seq_dir,
hr_seq,
data['frm_idx'],
to_bgr=True)
# compute metrics for the current sequence
if metric_calculator is not None:
gt_seq = data['gt'].numpy()
metric_calculator.compute_sequence_metrics(
data['seq_idx'], gt_seq, hr_seq)
# save/print results
if metric_calculator is not None:
seq_idx_lst = [
data['seq_idx'] for data in test_dataset
]
metric_calculator.gather(seq_idx_lst)
if opt['test'].get('save_json'):
# write results to a json file
json_path = osp.join(opt['test']['json_dir'],
f'{ds_name}_avg.json')
metric_calculator.save(model_idx,
json_path,
override=True)
else:
# print directly
metric_calculator.display()
model.load_state_dict(checkpoint['state_dict'])
# predict the characters
output = model.forward(tensor)
_, preds = torch.max(output, 1)
print(preds.item())
for p in np.array(preds.cpu()):
result.append(cat_to_class[model.class_to_idx[p]])
print(preds)
return result
if __name__ == '__main__':
if is_train:
model_ft, criterion, optimizer_ft, exp_lr_scheduler = models.define_model(
)
models = train.train_model(model_ft,
criterion,
optimizer_ft,
exp_lr_scheduler,
num_epochs=15)
if is_save:
torch.save(
{
'state_dict': model_ft.state_dict(),
'class_to_idx': model_ft.class_to_idx
}, os.path.join(model_path, 'cnnnet_win.pkl'))
else:
test.test()
from __future__ import absolute_import from __future__ import division from __future__ import print_function # import numpy as np # import os # import pickle # import tensorflow as tf # import time from data import bern_emb_data from models import define_model from args import parse_args from utils import make_dir args = parse_args() d = bern_emb_data(args.cs, args.ns, args.fpath, args.dynamic, args.n_epochs) dir_name = make_dir(d.name) m = define_model(args, d, dir_name) m.initialize_training() m.train_embeddings()
def train(args):
'''\
Training function.
Args:
args: namespace of arguments. Run 'artRecycle train --help' for info.
'''
# Model name and paths
model_name = '{}|{}'.format(*args.datasets)
model_path, log_path, logs_path = _prepare_directories(model_name,
resume=args.cont)
model_json = os.path.join(model_path, 'keras.json')
model_checkpoint = os.path.join(model_path, 'model')
# Summary writers
train_summary_writer = tf.summary.create_file_writer(
os.path.join(log_path, 'train'))
# Define datasets
image_shape = (300, 300, 3)
train_dataset, train_size = data.load_pair(*args.datasets,
'all',
shape=image_shape,
batch=args.batch)
train_dataset_it = iter(train_dataset)
few_samples = [data.load_few(name, 'all', image_shape, 1) \
for name in args.datasets]
# Define keras model
keras_model, model_layer = models.define_model(image_shape)
# Save keras model
keras_json = keras_model.to_json()
keras_json = json.dumps(json.loads(keras_json), indent=2)
with open(model_json, 'w') as f:
f.write(keras_json)
# Save TensorBoard graph
@tf.function
def tracing_model_ops(inputs):
return model_layer(inputs)
tf.summary.trace_on()
tracing_model_ops(next(train_dataset_it))
with train_summary_writer.as_default():
tf.summary.trace_export('Model', step=0)
# Resuming
if args.cont:
keras_model.load_weights(model_checkpoint)
print('> Weights loaded')
# Training steps
step_saver = CountersSaver(log_dir=logs_path, log_every=args.logs)
steps_per_epoch = int(train_size/args.batch) \
if not args.epoch_steps else args.epoch_steps
epochs = range(step_saver.epoch, args.epochs)
# Training tools
make_optmizer = lambda: tf.optimizers.Adam(args.rate)
trainer = models.Trainer(keras_model, make_optmizer, train_dataset_it)
tester = models.Tester(keras_model, train_dataset_it)
saver = CheckpointSaver(keras_model, model_checkpoint)
# Print job
print('> Training. Epochs:', epochs)
# Training loop
for epoch in epochs:
print('> Epoch', step_saver.epoch)
for epoch_step in range(steps_per_epoch):
print('> Step', step_saver.step, end='\r')
# Train step
output = trainer.step()
# Validation and log
if step_saver.step % args.logs == 0 or epoch_step == steps_per_epoch - 1:
print('\n> Validation')
# Evaluation
for i in range(args.val_steps):
tester.step()
train_metrics = tester.result()
# Log in console
print(' Train metrics:', train_metrics)
# Log in TensorBoard
with train_summary_writer.as_default():
for metric in train_metrics:
tf.summary.scalar(metric,
train_metrics[metric],
step=step_saver.step)
# Save weigths
loss = 0
for m in train_metrics.values():
loss += m
saved = saver.save(score=-loss)
if saved:
print('Weigths saved')
# Transform images for visualization
if args.images:
fake_A, fake_B, *_ = keras_model(few_samples)
fake_A_viz = image_unnormalize(fake_A)
fake_B_viz = image_unnormalize(fake_B)
# Log images
with train_summary_writer.as_default():
tf.summary.image('fake_A',
fake_A_viz,
step=step_saver.step)
tf.summary.image('fake_B',
fake_B_viz,
step=step_saver.step)
# End step
step_saver.new_step()
# End epoch
step_saver.new_epoch()
num_workers=threads)
test_loader = DataLoader(dataset,
batch_size=batch_size,
sampler=test_sampler,
shuffle=False,
num_workers=threads)
print('--training samples count:', len(train_indices))
print('--validation samples count:', len(val_indices))
print('--test samples count:', len(test_indices))
print('===> Loading model')
net = define_model(config['input_nc'],
config['output_nc'],
config['nfg'],
n_blocks=config['layers'],
gpu_id=device,
args=args).float()
# if torch.cuda.device_count() > 1:
# print("--using", torch.cuda.device_count(), "GPUs")
# net = nn.DataParallel(net)
# net.to(device)
# net_g.to(device)
optimizer = optim.Adam(net.parameters(),
lr=config['adam_lr'],
betas=(config['adam_b1'], config['adam_b2']))
net_scheduler = get_scheduler(optimizer, args)
criterionL1 = nn.L1Loss().to(device)
def train(opt):
# logging
logger = base_utils.get_logger('base')
logger.info('{} Options {}'.format('='*20, '='*20))
base_utils.print_options(opt, logger)
# create data loader
train_loader = create_dataloader(opt, dataset_idx='train')
# create downsampling kernels for BD degradation
kernel = data_utils.create_kernel(opt)
# create model
model = define_model(opt)
# training configs
total_sample = len(train_loader.dataset)
iter_per_epoch = len(train_loader)
total_iter = opt['train']['total_iter']
total_epoch = int(math.ceil(total_iter / iter_per_epoch))
curr_iter = opt['train']['start_iter']
test_freq = opt['test']['test_freq']
log_freq = opt['logger']['log_freq']
ckpt_freq = opt['logger']['ckpt_freq']
sigma_freq = opt['dataset']['degradation'].get('sigma_freq', 0)
sigma_inc = opt['dataset']['degradation'].get('sigma_inc', 0)
sigma_max = opt['dataset']['degradation'].get('sigma_max', 10)
logger.info('Number of training samples: {}'.format(total_sample))
logger.info('Total epochs needed: {} for {} iterations'.format(
total_epoch, total_iter))
print('device count:', torch.cuda.device_count())
# train
for epoch in range(total_epoch):
for data in tqdm(train_loader):
# update iter
curr_iter += 1
if curr_iter > total_iter:
logger.info('Finish training')
break
# update learning rate
model.update_learning_rate()
# prepare data
data = prepare_data(opt, data, kernel)
# train for a mini-batch
model.train(data)
# update running log
model.update_running_log()
# log
if log_freq > 0 and curr_iter % log_freq == 0:
# basic info
msg = '[epoch: {} | iter: {}'.format(epoch, curr_iter)
for lr_type, lr in model.get_current_learning_rate().items():
msg += ' | {}: {:.2e}'.format(lr_type, lr)
msg += '] '
# loss info
log_dict = model.get_running_log()
msg += ', '.join([
'{}: {:.3e}'.format(k, v) for k, v in log_dict.items()])
if opt['dataset']['degradation']['type'] == 'BD':
msg += ' | Sigma: {}'.format(opt['dataset']['degradation']['sigma'])
logger.info(msg)
# save model
if ckpt_freq > 0 and curr_iter % ckpt_freq == 0:
model.save(curr_iter)
# evaluate performance
if test_freq > 0 and curr_iter % test_freq == 0:
# setup model index
model_idx = 'G_iter{}'.format(curr_iter)
if opt['dataset']['degradation']['type'] == 'BD':
model_idx = model_idx + str(opt['dataset']['degradation']['sigma'])
# for each testset
for dataset_idx in sorted(opt['dataset'].keys()):
# use dataset with prefix `test`
if not dataset_idx.startswith('validate'):
continue
validate(opt, model, logger, dataset_idx, model_idx)
# schedule sigma
if opt['dataset']['degradation']['type'] == 'BD':
if sigma_freq > 0 and (epoch + 1) % sigma_freq == 0:
current_sigma = opt['dataset']['degradation']['sigma']
opt['dataset']['degradation']['sigma'] = min(current_sigma + sigma_inc, sigma_max)
kernel = data_utils.create_kernel(opt)
# __getitem__ in custom dataset class uses some crop that depends sigma
# it is crucial to change this cropsize accordingly if sigma is being changed
train_loader.dataset.change_cropsize(opt['dataset']['degradation']['sigma'])
print('kernel changed')
