[fix] fix unused tf_example6 files, fix uni-test

Author: ClarkChin08

examples/helloworld/tf_example6/dataset

@@ -84,7 +84,7 @@ def __init__(self, root, img_dir='val2017', \
         import os
         import numpy as np
         from pycocotools.coco import COCO
-        from lpot.metric.coco_label_map import category_map
+        from lpot.experimental.metric.coco_label_map import category_map
         self.image_list = []
         self.transform = transform
         self.filter = filter

examples/helloworld/tf_example6/lpot_workspace/tensorflow/mobilenet_v1/deploy.yaml

@@ -8,7 +8,7 @@
 
 sys.path.append('..')
 import lpot
-from lpot.experimental import common
+from lpot.experimental import Quantization, common
 
 def get_mlp_sym():
     data = mx.symbol.Variable('data')
@@ -102,8 +102,8 @@ def setUpClass(self):
         build_mxnet_kl()
         self.mlp_model = get_mlp_sym()
         self.conv_model = get_conv_sym()
-        self.quantizer_1 = lpot.Quantization("./mxnet.yaml")
-        self.quantizer_2 = lpot.Quantization("./mxnet_kl.yaml")
+        self.quantizer_1 = Quantization("./mxnet.yaml")
+        self.quantizer_2 = Quantization("./mxnet_kl.yaml")
         framework_specific_info = dict()
         framework_specific_info['q_dataloader'] = None
         self.adaptor = lpot.adaptor.mxnet.MxNetAdaptor(framework_specific_info)

examples/helloworld/tf_example6/small_dataset

@@ -44,8 +44,8 @@ def test_register_metric_postprocess(self):
         images = np.expand_dims(resize_image, axis=0)
         labels = [768]
         from lpot import Benchmark, Quantization
-        from lpot.data.transforms.imagenet_transform import LabelShift
-        from lpot.metric.metric import TensorflowTopK
+        from lpot.experimental.data.transforms.imagenet_transform import LabelShift
+        from lpot.experimental.metric.metric import TensorflowTopK
 
         evaluator = Benchmark('fake_yaml.yaml')
 
@@ -68,7 +68,6 @@ def test_register_metric_postprocess(self):
         acc, batch_size, result_list = result['accuracy']
         self.assertEqual(acc, 0.0)
 
-        
 
 if __name__ == "__main__":
     unittest.main()

lpot/experimental/data/datasets/coco_dataset.py

@@ -43,6 +43,5 @@ def test_convert_internal_patterns(self):
         self.assertEqual([['AvgPool']] in internal_patterns, True)
         self.assertEqual([['MatMul'], ('BiasAdd',), ('Relu',)] in internal_patterns, True)
 
-
 if __name__ == '__main__':
     unittest.main()

test/test_adaptor_mxnet.py

@@ -766,7 +766,7 @@ def setUpClass(cls):
         cls.vocab_file = 'bert/vocab.txt'
 
     def testSquadV1PostAndF1(self):
-        from lpot.data.transforms.transform import SquadV1PostTransform
+        from lpot.experimental.data.transforms.transform import SquadV1PostTransform
         squad_post = SquadV1PostTransform(self.label_file, self.vocab_file) 
         unique_ids=np.arange(1000000000, 1000010833) 
         start_logits=np.ones((10833, 384), np.float32)
@@ -781,21 +781,21 @@ def get_labels(label_file):
 
         labels = get_labels(self.label_file)
         preds, labels = squad_post((preds, labels))
-        from lpot.metric.metric import SquadF1
+        from lpot.experimental.metric.metric import SquadF1
         squad_metric = SquadF1()
         squad_metric.update(preds, labels)
         result = squad_metric.result()
         self.assertEqual(round(result, 2), 1.92)
         squad_metric.reset()
 
     def testBertDataLoader(self):
-        from lpot.data.datasets.bert_dataset import TensorflowBertDataset
+        from lpot.experimental.data.datasets.bert_dataset import TensorflowBertDataset
         fake_record='fake.tf_record'
         bert_dataset = TensorflowBertDataset(fake_record, self.label_file) 
         self.assertEqual(len(bert_dataset), 1)
         get_record, _ = bert_dataset[0]
         self.assertEqual(fake_record, get_record)
-        from lpot.data.dataloaders.tensorflow_dataloader import TensorflowDataLoader
+        from lpot.experimental.data.dataloaders.tensorflow_dataloader import TensorflowDataLoader
         bert_dataloader = TensorflowDataLoader(bert_dataset, batch_size=1)
         iterator = iter(bert_dataloader)
         (get_record, batch_size), _ = next(iterator)