load_ckpt debug

model_zoo/official/cv/mobilenetv2/eval.py

@@ -46,6 +46,10 @@ if __name__ == '__main__':
 
     dataset = create_dataset(dataset_path=args_opt.dataset_path, do_train=False, config=config)
     step_size = dataset.get_dataset_size()
+    if step_size == 0:
+        raise ValueError("The step_size of dataset is zero. Check if the images count of train dataset is more \
+            than batch_size in config.py")
+
     net.set_train(False)
 
     loss = nn.SoftmaxCrossEntropyWithLogits(sparse=True, reduction='mean')

model_zoo/official/cv/mobilenetv2/src/dataset.py

@@ -16,7 +16,6 @@
 create train or eval dataset.
 """
 import os
-from tqdm import tqdm
 import numpy as np
 
 from mindspore import Tensor
@@ -109,19 +108,20 @@ def extract_features(net, dataset_path, config):
                              config=config,
                              repeat_num=1)
     step_size = dataset.get_dataset_size()
-    pbar = tqdm(list(dataset.create_dict_iterator()))
+    if step_size == 0:
+        raise ValueError("The step_size of dataset is zero. Check if the images count of train dataset is more \
+            than batch_size in config.py")
+
     model = Model(net)
-    i = 0
+    for i, data in enumerate(dataset.create_dict_iterator()):
-    for data in pbar:
         features_path = os.path.join(features_folder, f"feature_{i}.npy")
         label_path = os.path.join(features_folder, f"label_{i}.npy")
-        if not (os.path.exists(features_path) and os.path.exists(label_path)):
+        if not os.path.exists(features_path) or not os.path.exists(label_path):
             image = data["image"]
             label = data["label"]
             features = model.predict(Tensor(image))
             np.save(features_path, features.asnumpy())
             np.save(label_path, label)
-        pbar.set_description("Process dataset batch: %d" % (i + 1))
+        print(f"Complete the batch {i}/{step_size}")
-        i += 1
 
     return step_size

model_zoo/official/cv/mobilenetv2/src/mobilenetV2.py

@@ -330,8 +330,12 @@ class MobileNetV2(nn.Cell):
     MobileNetV2 architecture.
 
     Args:
-        backbone(nn.Cell):
+        class_num (int): number of classes.
-        head(nn.Cell):
+        width_mult (int): Channels multiplier for round to 8/16 and others. Default is 1.
+        has_dropout (bool): Is dropout used. Default is false
+        inverted_residual_setting (list): Inverted residual settings. Default is None
+        round_nearest (list): Channel round to . Default is 8
+
     Returns:
         Tensor, output tensor.
 
@@ -355,14 +359,11 @@ class MobileNetV2(nn.Cell):
 
 class MobileNetV2Combine(nn.Cell):
     """
-    MobileNetV2 architecture.
+    MobileNetV2Combine architecture.
 
     Args:
-        class_num (Cell): number of classes.
+        backbone(Cell): The features extract layers.
-        width_mult (int): Channels multiplier for round to 8/16 and others. Default is 1.
+        head(Cell): The fully connected layer.
-        has_dropout (bool): Is dropout used. Default is false
-        inverted_residual_setting (list): Inverted residual settings. Default is None
-        round_nearest (list): Channel round to . Default is 8
     Returns:
         Tensor, output tensor.
 
@@ -371,7 +372,7 @@ class MobileNetV2Combine(nn.Cell):
     """
 
     def __init__(self, backbone, head):
-        super(MobileNetV2Combine, self).__init__()
+        super(MobileNetV2Combine, self).__init__(auto_prefix=False)
         self.backbone = backbone
         self.head = head
 

model_zoo/official/cv/mobilenetv2/train.py

@@ -62,6 +62,9 @@ if __name__ == '__main__':
             raise ValueError("Currently, CPU only support \"incremental_learn\", not \"fine_tune\" or \"train\".")
         dataset = create_dataset(dataset_path=args_opt.dataset_path, do_train=True, config=config)
         step_size = dataset.get_dataset_size()
+        if step_size == 0:
+            raise ValueError("The step_size of dataset is zero. Check if the images count of train dataset is more \
+                than batch_size in config.py")
 
     # Currently, only Ascend support switch precision.
     switch_precision(net, mstype.float16, config)
@@ -108,9 +111,8 @@ if __name__ == '__main__':
                 losses.append(network(feature, label).asnumpy())
             epoch_mseconds = (time.time()-epoch_start) * 1000
             per_step_mseconds = epoch_mseconds / step_size
-            print("\r epoch[{}], iter[{}] cost: {:5.3f}, per step time: {:5.3f}, avg loss: {:5.3f}"\
+            print("epoch[{}], iter[{}] cost: {:5.3f}, per step time: {:5.3f}, avg loss: {:5.3f}"\
-            .format(epoch + 1, step_size, epoch_mseconds, per_step_mseconds, np.mean(np.array(losses))), \
+            .format(epoch + 1, step_size, epoch_mseconds, per_step_mseconds, np.mean(np.array(losses))))
-                end="")
             if (epoch + 1) % config.save_checkpoint_epochs == 0:
                 _exec_save_checkpoint(network, os.path.join(config.save_checkpoint_path, \
                     f"mobilenetv2_head_{epoch+1}.ckpt"))