move model to cuda in tests, add comment (#2384)

Author: joelgrus

allennlp/tests/training/trainer_test.py

@@ -91,6 +91,7 @@ def test_trainer_can_run(self):
 
     @pytest.mark.skipif(not torch.cuda.is_available(), reason="No CUDA device registered.")
     def test_trainer_can_run_cuda(self):
+        self.model.cuda()
         trainer = Trainer(self.model, self.optimizer,
                           self.iterator, self.instances, num_epochs=2,
                           cuda_device=0)
@@ -99,7 +100,7 @@ def test_trainer_can_run_cuda(self):
     @pytest.mark.skipif(torch.cuda.device_count() < 2,
                         reason="Need multiple GPUs.")
     def test_trainer_can_run_multiple_gpu(self):
-
+        self.model.cuda()
         class MetaDataCheckWrapper(Model):
             """
             Checks that the metadata field has been correctly split across the batch dimension

allennlp/training/trainer.py

@@ -69,6 +69,10 @@ def __init__(self,
             An AllenNLP model to be optimized. Pytorch Modules can also be optimized if
             their ``forward`` method returns a dictionary with a "loss" key, containing a
             scalar tensor representing the loss function to be optimized.
+
+            If you are training your model using GPUs, your model should already be
+            on the correct device. (If you use `Trainer.from_params` this will be
+            handled for you.)
         optimizer : ``torch.nn.Optimizer``, required.
             An instance of a Pytorch Optimizer, instantiated with the parameters of the
             model to be optimized.