def render():
model = torch.load_model('mymodel')
# transformations
testset = datasets.CIFAR10('./data-test', download=True, train=False, transform=transform)
# testloader
for _, data in enumerate(testloader, 0):
inputs, labels = data
outputs = model(inputs)
def predict(data, code='000001', folder='../dlmodels'):
model_file = folder + '/' + 'model_' + code + '.ptm'
model = load_model(model_file)
model.eval()
dataset = torch.FloatTensor(data)
x = Variable(dataset)
outputs = model(x)
outputs = outputs.data.numpy()
tmpvalues = np.append(data, outputs)
scaler = make_scaler(code=code)
tmpvalues = scaler.inverse_transform(
np.reshape(tmpvalues, (1, tmpvalues.shape[0])))
tmpvalues = tmpvalues[0]
outputs = tmpvalues[feature * timestep:]
return format_outputs(outputs)
def load(self, model_name):
self.modelF.load_state_dict(load_model(os.path.join('model', model_name)))
def load(self, name):
self.model = load_model(name)
def load_model():
return torch.load_model(file2save)
def load(self, name):
self.model = load_model(name)
self.target_model = load_model(name)