fix crnn postprocess error

model_zoo/official/cv/crnn/README.md

@@ -39,6 +39,11 @@ CRNN was a neural network for image based sequence recognition and its Applicati
 
 CRNN use a vgg16 structure for feature extraction, the appending with two-layer bidirectional LSTM, finally use CTC to calculate loss. See src/crnn.py for details.
 
+We provide 2 versions of network using different ways to transfer the hidden size to class numbers. You could choose different version by modifying the `model_version` in config yaml.
+
+- V1 using an full connection after the RNN parts.
+- V2 change the output feature size of the last RNN, to output a feature with the same size of class numbers.
+
 ## [Dataset](#content)
 
 Note that you can run the scripts based on the dataset mentioned in original paper or widely used in relevant domain/network architecture. In the following sections, we will introduce how to run the scripts using the related dataset below.

model_zoo/official/cv/crnn/default_config.yaml

@@ -8,14 +8,14 @@ checkpoint_url: ""
 data_path: "/cache/data"
 output_path: "/cache/train"
 load_path: "/cache/checkpoint_path"
-device_target: "GPU"
+device_target: "Ascend"
 enable_profiling: False
 
 # ======================================================================================
 # common options
 run_distribute: False
 model: "lowercase"
-model_version: "V2"
+model_version: "V1"
 
 # ======================================================================================
 # Training options
@@ -64,8 +64,8 @@ file_format: "MINDIR"
 
 # ======================================================================================
 #postprocess
-ann_file: True
-result_path: True
+ann_file: ""
+result_path: ""
 dataset: "ic03"
 
 ---