added some code to test reported issues

Signed-off-by: Jules Damji <[email protected]>

Author: dmatrix

py/mlflow/mlflow_issues/issue_slack_tf_model.py

@@ -0,0 +1,41 @@
+import numpy as np
+import mlflow
+
+QADict = {'What is your full name?': 'My name is King George.',
+          'Have you taken your medication?': "I don't think so."}
+
+
+class convEngine(mlflow.pyfunc.PythonModel):
+
+    def __init__(self, QADict, model):
+        self.model = model
+        self.Qlist = list(QADict.keys())
+        self.Qanswers = self.embed(self.Qlist)
+        self.QADict = QADict
+
+    def embed(self, input):
+        return self.model(input)
+
+    def predict(self, model_input_string):
+        qasked = self.embed([model_input_string])
+        corr = np.inner(qasked, self.Qanswers)
+        index_max = np.argmax(corr)
+
+        return self.QADict[self.Qlist[index_max]]
+
+
+if __name__ == '__main__':
+
+    import tensorflow_hub as hub
+    import cloudpickle
+
+    module_url = "https://tfhub.dev/google/universal-sentence-encoder-large/5"
+    model = hub.load(module_url)
+    model_path = "/tmp/tf_model"
+
+    with open(model_path, "w+") as f:
+        cloudpickle.dump(model, f)
+    """
+    convModel = convEngine(QADict, model)
+    mlflow.pyfunc.save_model(path=model_path, python_model=convModel)
+    """

py/mlflow/mlflow_issues/save_model.py

@@ -0,0 +1,22 @@
+import mlflow.sklearn
+from sklearn.datasets import load_iris
+from sklearn import tree
+
+if __name__ == '__main__':
+
+    iris = load_iris()
+    sk_model = tree.DecisionTreeClassifier()
+    sk_model = sk_model.fit(iris.data, iris.target)
+
+    sk_path = 'saved_models'
+    mlflow.set_tracking_uri('sqlite:///mlruns.db')
+    with mlflow.start_run() as run:
+        mlflow.log_metric('m', 1.5)
+        mlflow.sklearn.save_model(sk_model, sk_path)
+
+    # Register model
+    model_uri = "runs:/{}/model".format(run.info.run_id)
+    registered_model_name = "RegisterSavedModel"
+    mv = mlflow.register_model(model_uri, registered_model_name)
+    print("Name: {}".format(mv.name))
+    print("Version: {}".format(mv.version))

pytorch/mlflow/pytorch_autolog_mlflow_example.py

@@ -53,36 +53,43 @@ def forward(self, x):
 
         return x
 
+    @staticmethod
+    def add_model_specific_args(parent_parser):
 
-if __name__ == "__main__":
-    parser = ArgumentParser(description="PyTorch Autolog Mnist Example")
+        parser = ArgumentParser(parents=[parent_parser], add_help=False)
 
-    # Add trainer specific arguments
+        # Add trainer specific arguments
+        parser.add_argument(
+            "--tracking_uri", type=str, default="http://localhost:5000/", help="mlflow tracking uri"
+        )
+        parser.add_argument(
+            "--max_epochs", type=int, default=20, help="number of epochs to run (default: 20)"
+        )
+        parser.add_argument(
+            "--gpus", type=int, default=0, help="Number of gpus - by default runs on CPU"
+        )
+        parser.add_argument(
+            "--accelerator",
+            type=str,
+            default=None,
+            help="accelerator - (default: None)",
+        )
+        return parser
 
-    parser.add_argument(
-        "--tracking_uri", type=str, default="http://localhost:5000/", help="mlflow tracking uri"
-    )
-    parser.add_argument(
-        "--max_epochs", type=int, default=20, help="number of epochs to run (default: 20)"
-    )
-    parser.add_argument(
-        "--gpus", type=int, default=0, help="Number of gpus - by default runs on CPU"
-    )
-    parser.add_argument(
-        "--accelerator",
-        type=str,
-        default=None,
-        help="accelerator - (default: None)",
-    )
-    parser = LightningMNISTClassifier.add_model_specific_args(parent_parser=parser)
+
+if __name__ == "__main__":
+    parent_parser = ArgumentParser(description="PyTorch Autolog Mnist Example")
+
+    parser = LightningMNISTClassifier.add_model_specific_args(parent_parser=parent_parser)
 
     mlflow.pytorch.autolog()  # just add this line and your Autologging should work!
 
     args = parser.parse_args()
 
     args = parser.parse_args()
     dict_args = vars(args)
-    #mlflow.set_tracking_uri(dict_args['tracking_uri'])
+
+    # mlflow.set_tracking_uri(dict_args['tracking_uri'])
 
     model = LightningMNISTClassifier(**dict_args)
     early_stopping = EarlyStopping(monitor="val_loss", mode="min", verbose=True)
@@ -93,15 +100,15 @@ def forward(self, x):
         verbose=True,
         monitor="val_loss",
         mode="min",
-        prefix="",
+        prefix=""
     )
     lr_logger = LearningRateMonitor()
 
     trainer = pl.Trainer.from_argparse_args(
         args,
         callbacks=[lr_logger, early_stopping],
         checkpoint_callback=checkpoint_callback,
-        train_percent_check=0.1,
+        # train_percent_check=0.1,
     )
     trainer.fit(model)
     trainer.test()

pytorch/tutorials/13_feed_forward.ipynb

@@ -277,7 +277,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.7.5"
+   "version": "3.7.4"
   }
  },
  "nbformat": 4,

pytorch/tutorials/14_cnn.ipynb

@@ -245,7 +245,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.7.5"
+   "version": "3.7.4"
   }
  },
  "nbformat": 4,

pytorch/tutorials/18_custom_dataset.ipynb

@@ -904,7 +904,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.7.5"
+   "version": "3.7.4"
   }
  },
  "nbformat": 4,