raise ValueError('Unknown method')
model = model.cuda()
# load model
start_epoch = params.start_epoch
stop_epoch = params.stop_epoch
if params.resume != '':
resume_file = get_resume_file(
'%s/checkpoints/%s' % (params.save_dir, params.resume),
params.resume_epoch)
if resume_file is not None:
tmp = torch.load(resume_file)
start_epoch = tmp['epoch'] + 1
model.load_state_dict(tmp['state'])
print(' resume the training with at {} epoch (model file {})'.
format(start_epoch, params.resume))
elif 'baseline' not in params.method:
if params.warmup == 'gg3b0':
raise Exception(
'Must provide the pre-trained feature encoder file using --warmup option!'
)
state = load_warmup_state(
'%s/checkpoints/%s' % (params.save_dir, params.warmup),
params.method)
model.feature.load_state_dict(state, strict=False)
# training
print('\n--- start the training ---')
model = train(base_loader, val_loader, model, start_epoch, stop_epoch,
params)
model.cuda()
# resume training
start_epoch = params.start_epoch
stop_epoch = params.stop_epoch
if params.resume != '':
resume_file = get_resume_file(
'%s/checkpoints/%s' % (params.save_dir, params.resume),
params.resume_epoch)
if resume_file is not None:
start_epoch = model.resume(resume_file)
print(' resume the training with at {} epoch (model file {})'.
format(start_epoch, params.resume))
else:
raise ValueError('No resume file')
# load pre-trained feature encoder
else:
if params.warmup == 'gg3b0':
raise Exception(
'Must provide pre-trained feature-encoder file using --warmup option!'
)
model.model.feature.load_state_dict(load_warmup_state(
'%s/checkpoints/%s' % (params.save_dir, params.warmup),
params.method),
strict=False)
# training
print('\n--- start the training ---')
train(base_datamgr, datasets, val_loader, model, start_epoch, stop_epoch,
params)
if torch.cuda.is_available():
model = model.cuda()
# load model
start_epoch = params.start_epoch #0
stop_epoch = params.stop_epoch #400
if params.resume != '':
resume_file = get_resume_file(
'%s/checkpoints/%s' % (params.save_dir, params.resume),
params.resume_epoch)
if resume_file is not None:
tmp = torch.load(resume_file)
start_epoch = tmp['epoch'] + 1
model.load_state_dict(tmp['state'])
print(' resume the training with at {} epoch (model file {})'.
format(start_epoch, params.resume))
elif 'baseline' not in params.method:
if params.warmup == 'gg3b0':
raise Exception(
'Must provide the pre-trained feature encoder file using --warmup option!'
)
state = load_warmup_state(
'%s/checkpoints/%s' % (params.save_dir, params.warmup),
params.method) # modify feature extractor paras
model.feature.load_state_dict(state, strict=False)
# training
print('\n--- start the training ---')
model = train(base_loader, val_loader, model, start_epoch, stop_epoch,
params)
