def evaluate(model, dataloader, epoch, writer, logger, data_name='val'):
save_root = os.path.join(opt.result_dir, opt.tag, str(epoch), data_name)
utils.try_make_dir(save_root)
total_psnr = 0.0
total_ssim = 0.0
ct_num = 0
# print('Start testing ' + tag + '...')
for i, sample in enumerate(dataloader):
utils.progress_bar(i, len(dataloader), 'Eva... ')
path = sample['path']
with torch.no_grad():
recovered = model(sample)
if data_name == 'val':
label = sample['label']
label = tensor2im(label)
recovered = tensor2im(recovered)
ct_num += 1
total_psnr += psnr(recovered, label, data_range=255)
total_ssim += ski_ssim(recovered,
label,
data_range=255,
multichannel=True)
save_dst = os.path.join(save_root,
utils.get_file_name(path[0]) + '.png')
Image.fromarray(recovered).save(save_dst)
elif data_name == 'test':
pass
else:
raise Exception('Unknown dataset name: %s.' % data_name)
# 保存结果
save_dst = os.path.join(save_root,
utils.get_file_name(path[0]) + '.png')
Image.fromarray(recovered).save(save_dst)
if data_name == 'val':
ave_psnr = total_psnr / float(ct_num)
ave_ssim = total_ssim / float(ct_num)
# write_loss(writer, f'val/{data_name}', 'psnr', total_psnr / float(ct_num), epochs)
logger.info(f'Eva({data_name}) epoch {epoch}, psnr: {ave_psnr}.')
logger.info(f'Eva({data_name}) epoch {epoch}, ssim: {ave_ssim}.')
return f'{ave_ssim: .3f}'
else:
return ''
def evaluate(model, dataloader, epochs, writer, logger, data_name='val'):
save_root = os.path.join(opt.result_dir, opt.tag, str(epochs), data_name)
utils.try_make_dir(save_root)
correct = 0
ct_num = 0
counts = defaultdict(int)
corrects = defaultdict(int)
# print('Start testing ' + tag + '...')
for i, data in enumerate(dataloader):
if data_name == 'val':
input, label, path = data['input'], data['label'], data['path']
utils.progress_bar(i, len(dataloader), 'Eva... ')
# ct_num += 1
with torch.no_grad():
img_var = Variable(input,
requires_grad=False).to(device=opt.device)
label_var = Variable(label,
requires_grad=False).to(device=opt.device)
predicted = model(img_var)
_, predicted = torch.max(predicted, 1)
ct_num += label.size(0)
correct += (predicted == label_var).sum().item()
for idx, l in enumerate(label):
l = l.item()
counts[l] += 1
p = predicted[idx].item()
if p == l:
corrects[l] += 1
elif data_name == 'test':
pass
else:
raise Exception('Unknown dataset name: %s.' % data_name)
if data_name == 'val':
# write_loss(writer, 'val/%s' % data_name, 'psnr', ave_psnr / float(ct_num), epochs)
acc = 0.
for k in counts:
acc += corrects[k] / counts[k]
err = 1 - acc / len(counts)
err = err + 0.03 # 线上错误率比线下高0.03
# logger.info('Eva(%s) epoch %d ,' % (data_name, epochs) + 'Accuracy: ' + str(correct / float(ct_num)) + '.')
logger.info('Eva(%s) epoch %d ,' % (data_name, epochs) + 'Err: ' +
str(err) + '.')
# ipdb.set_trace()
return str(round(correct / float(ct_num), 3))
else:
return ''
from network import get_model
from eval import evaluate
from utils import *
from mscv.summary import create_summary_writer, write_meters_loss, write_image
import misc_utils as utils
# 初始化
with torch.no_grad():
# 初始化路径
save_root = os.path.join(opt.checkpoint_dir, opt.tag)
log_root = os.path.join(opt.log_dir, opt.tag)
utils.try_make_dir(save_root)
utils.try_make_dir(log_root)
train_dataloader = dl.train_dataloader
val_dataloader = dl.val_dataloader
# 初始化日志
logger = init_log(training=True)
# 初始化模型
Model = get_model(opt.model)
model = Model(opt)
# 暂时还不支持多GPU
# if len(opt.gpu_ids):
# model = torch.nn.DataParallel(model, device_ids=opt.gpu_ids)
