def get_model():
print('[+] loading model... ', end='', flush=True)
model = models.densenet201_finetune(num_classes=NB_CLASSES, drop_rate=0.20)
if use_gpu:
model.cuda()
print('done')
return model
def get_model():
print('[+] loading model... ', end='', flush=True)
model = models.densenet201_finetune(NB_CLASSES)
if use_gpu:
model.cuda()
print('done')
return model
def get_model(name):
print('[+] loading model... ', end='', flush=True)
if name == 'densenet201':
model = models.densenet201_finetune(NB_CLASSES)
elif name == 'inceptionresnetv2':
model = models.inceptionresnetv2_finetune(NB_CLASSES)
if name == 'senet154':
model = models.senet154_finetune(NB_CLASSES)
if name == 'nasnetlarge':
model = models.nasnetlarge_finetune(NB_CLASSES)
if name == 'inceptionv4':
model = models.inceptionv4_finetune(NB_CLASSES)
if name == 'se_resnext101_32x4d':
model = models.se_resnext101_32x4d_finetune(NB_CLASSES)
if use_gpu:
model.cuda()
print('done')
return model
def get_model(model_name):
print('Loading model: %s' % (model_name))
if model_name.startswith("densenet"):
model = models.densenet201_finetune(NB_CLASSES)
elif model_name.startswith("squeezenet"):
model = models.squeezenet11_finetune(NB_CLASSES)
elif model_name.startswith("resnet"):
model = models.resnet152_finetune(NB_CLASSES)
else:
print("Error: Model not found!")
exit(-1)
# Multi-GPU scaling
if torch.cuda.device_count() > 1:
print("Parallelizing model over %d GPUs!" % (torch.cuda.device_count()))
model = nn.DataParallel(model)
model.to(device)
print('Model loaded successfully!')
return model