[GitHub] [beam] AnandInguva commented on a diff in pull request #26683: [Python] Added load model option args for PyTorch Model Handler

["AnandInguva (via GitHub)" <gi...@apache.org>]

AnandInguva commented on code in PR #26683:
URL: https://github.com/apache/beam/pull/26683#discussion_r1193880474


##########
sdks/python/apache_beam/ml/inference/pytorch_inference_test.py:
##########
@@ -893,5 +894,55 @@ def test_env_vars_set_correctly_keyed_tensor_handler(self):
       self.assertTrue((os.environ['FOO']) == 'bar')
 
 
+@pytest.mark.uses_pytorch
+class PytorchInferenceTestWithMocks(unittest.TestCase):
+  def setUp(self):
+    self._load_model = pytorch_inference._load_model
+    pytorch_inference._load_model = unittest.mock.MagicMock(
+        return_value=("model", "device"))
+    self.tmpdir = tempfile.mkdtemp()
+    self.state_dict = OrderedDict([('linear.weight', torch.Tensor([[2.0, 3]])),
+                                   ('linear.bias', torch.Tensor([0.5]))])
+    self.torch_path = os.path.join(self.tmpdir, 'torch_model.pt')
+    torch.save(self.state_dict, self.torch_path)
+    self.model_params = {'input_dim': 2, 'output_dim': 1}
+
+  def tearDown(self):
+    pytorch_inference._load_model = self._load_model
+    shutil.rmtree(self.tmpdir)
+
+  def test_load_model_args_tensor(self):
+    load_model_args = {'weights_only': True}
+    model_handler = PytorchModelHandlerTensor(
+        state_dict_path=self.torch_path,
+        model_class=PytorchLinearRegression,
+        model_params=self.model_params,
+        load_model_args=load_model_args)
+    model_handler.load_model()
+    pytorch_inference._load_model.assert_called_with(
+        PytorchLinearRegression,
+        self.torch_path,

Review Comment:
   Can we pass the args using keywords? just for readability purpose



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscribe@beam.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org