// TODO(Mark): Generalise how the array is transformed into a variable after settling the data API.
// Add a batch dimension by unsqueezing, because pytorch
// doesn"t support inputs without one.
array_input = {"tokens": torch.autograd.Variable(torch.LongTensor(array_input["tokens"])).unsqueeze(0)}
output_dict = self.forward(tokens=array_input)
// Remove batch dimension, as we only had one input.
After Change
// TODO(Mark): Make the data API always return tensors with batch dimensions at every abstraction level.
// Add a batch dimension by unsqueezing, because pytorch doesn"t support inputs without one.
model_input = arrays_to_variables(array_input)
model_input["tokens"].data.unsqueeze_(0)
output_dict = self.forward(tokens=model_input)
// Remove batch dimension, as we only had one input.
predictions = output_dict["class_probabilities"].data.squeeze(0)
