!151 fix calculation of lineage dataset_num, get model name

Merge pull request !151 from luopengting/lineage_fix

mindinsight/lineagemgr/collection/model/model_lineage.py

@@ -37,7 +37,7 @@ from mindinsight.lineagemgr.collection.model.base import Metadata
 try:
     from mindspore.common.tensor import Tensor
     from mindspore.train.callback import Callback, RunContext, ModelCheckpoint, SummaryStep
-    from mindspore.nn import Cell, Optimizer, WithLossCell, TrainOneStepWithLossScaleCell
+    from mindspore.nn import Cell, Optimizer
     from mindspore.nn.loss.loss import _Loss
     from mindspore.dataset.engine import Dataset, MindDataset
     import mindspore.dataset as ds
@@ -412,27 +412,28 @@ class AnalyzeObject:
         Returns:
             str, the name of the backbone network.
         """
-        with_loss_cell = False
-        backbone = None
+        backbone_name = None
+        has_network = False
+        network_key = 'network'
+        backbone_key = '_backbone'
+
         net_args = vars(network) if network else {}
         net_cell = net_args.get('_cells') if net_args else {}
 
-        for _, value in net_cell.items():
-            if isinstance(value, WithLossCell):
-                backbone = getattr(value, '_backbone')
-                with_loss_cell = True
+        for key, value in net_cell.items():
+            if key == network_key:
+                network = value
+                has_network = True
                 break
 
-        if with_loss_cell:
-            backbone_name = type(backbone).__name__ \
-                if backbone else None
-        elif isinstance(network, TrainOneStepWithLossScaleCell):
-            backbone = getattr(network, 'network')
-            backbone_name = type(backbone).__name__ \
-                if backbone else None
-        else:
-            backbone_name = type(network).__name__ \
-                if network else None
+        if has_network:
+            while hasattr(network, network_key):
+                network = getattr(network, network_key)
+            if hasattr(network, backbone_key):
+                backbone = getattr(network, backbone_key)
+                backbone_name = type(backbone).__name__
+        elif network is not None:
+            backbone_name = type(network).__name__
         return backbone_name
 
     @staticmethod
@@ -489,26 +490,24 @@ class AnalyzeObject:
         Returns:
             dict, the lineage metadata.
         """
-        dataset_batch_size = dataset.get_dataset_size()
-        if dataset_batch_size is not None:
-            validate_int_params(dataset_batch_size, 'dataset_batch_size')
-        log.debug('dataset_batch_size: %d', dataset_batch_size)
+        batch_num = dataset.get_dataset_size()
+        batch_size = dataset.get_batch_size()
+        if batch_num is not None:
+            validate_int_params(batch_num, 'dataset_batch_num')
+            validate_int_params(batch_num, 'dataset_batch_size')
+        log.debug('dataset_batch_num: %d', batch_num)
+        log.debug('dataset_batch_size: %d', batch_size)
         dataset_path = AnalyzeObject.get_dataset_path_wrapped(dataset)
         if dataset_path:
             dataset_path = '/'.join(dataset_path.split('/')[:-1])
 
-        step_num = lineage_dict.get('step_num')
-        validate_int_params(step_num, 'step_num')
-        log.debug('step_num: %d', step_num)
-
+        dataset_size = int(batch_num * batch_size)
         if dataset_type == 'train':
             lineage_dict[Metadata.train_dataset_path] = dataset_path
-            epoch = lineage_dict.get('epoch')
-            train_dataset_size = dataset_batch_size * (step_num / epoch)
-            lineage_dict[Metadata.train_dataset_size] = int(train_dataset_size)
+            lineage_dict[Metadata.train_dataset_size] = dataset_size
         elif dataset_type == 'valid':
             lineage_dict[Metadata.valid_dataset_path] = dataset_path
-            lineage_dict[Metadata.valid_dataset_size] = dataset_batch_size * step_num
+            lineage_dict[Metadata.valid_dataset_size] = dataset_size
 
         return lineage_dict
 

mindinsight/lineagemgr/querier/query_model.py

@@ -82,7 +82,6 @@ class LineageObj:
         self._lineage_info = {
             self._name_summary_dir: summary_dir
         }
-        self._filtration_result = None
         self._init_lineage()
         self.parse_and_update_lineage(**kwargs)
 

tests/st/func/lineagemgr/api/test_model_api.py

@@ -50,10 +50,10 @@ LINEAGE_INFO_RUN1 = {
         'network': 'ResNet'
     },
     'train_dataset': {
-        'train_dataset_size': 731
+        'train_dataset_size': 1024
     },
     'valid_dataset': {
-        'valid_dataset_size': 10240
+        'valid_dataset_size': 1024
     },
     'model': {
         'path': '{"ckpt": "'
@@ -89,9 +89,9 @@ LINEAGE_FILTRATION_RUN1 = {
     'model_lineage': {
         'loss_function': 'SoftmaxCrossEntropyWithLogits',
         'train_dataset_path': None,
-        'train_dataset_count': 731,
+        'train_dataset_count': 1024,
         'test_dataset_path': None,
-        'test_dataset_count': 10240,
+        'test_dataset_count': 1024,
         'user_defined': {},
         'network': 'ResNet',
         'optimizer': 'Momentum',
@@ -115,7 +115,7 @@ LINEAGE_FILTRATION_RUN2 = {
         'train_dataset_path': None,
         'train_dataset_count': 1024,
         'test_dataset_path': None,
-        'test_dataset_count': 10240,
+        'test_dataset_count': 1024,
         'user_defined': {},
         'network': "ResNet",
         'optimizer': "Momentum",

tests/ut/lineagemgr/collection/model/test_model_lineage.py

@@ -334,12 +334,14 @@ class TestAnalyzer(TestCase):
         )
         res1 = self.analyzer.analyze_dataset(dataset, {'step_num': 10, 'epoch': 2}, 'train')
         res2 = self.analyzer.analyze_dataset(dataset, {'step_num': 5}, 'valid')
+
+        # batch_size is mocked as 32.
         assert res1 == {'step_num': 10,
                         'train_dataset_path': '/path/to',
-                        'train_dataset_size': 50,
+                        'train_dataset_size': 320,
                         'epoch': 2}
         assert res2 == {'step_num': 5, 'valid_dataset_path': '/path/to',
-                        'valid_dataset_size': 50}
+                        'valid_dataset_size': 320}
 
     def test_get_dataset_path_dataset(self):
         """Test get_dataset_path method with Dataset."""

tests/utils/mindspore/dataset/engine/datasets.py

@@ -27,6 +27,10 @@ class Dataset:
         """Mocked get_dataset_size."""
         return self.dataset_size
 
+    def get_batch_size(self):
+        """Mocked get_batch_size"""
+        return 32
+
 
 class MindDataset(Dataset):
     """Mock the MindSpore MindDataset class."""