i need help - predict with openvino

[staebchen0]

Hi, after many tries I finally got the code up and running. Now it just depends on one mistake ...

My question: I trained the original hdf5 model with MobileNet. My model has 3 classes

According to your instructions, I have converted my model into an xml and am. However, get an error in the predict

In your example "inception_v3" is called ...

from tensorflow.keras.applications.inception_v3 import decode_predictions
print ('Predicted:', decode_predictions (res, top = 3) [0])

decode_predictions expects a batch of predictions (i.e. a 2D array of shape (samples, 1000)). Found array with shape: (1, 2)

I am trying to speed up my prediction time with the help of openvino (monitoring live cam) For the prediction I currently need about 15 milliseconds on average. Can you improve that with openvino at all?

Here is my timekeeping:

[INFO] Start Predict..Sat Apr 4 08:50:10 2020 [INFO] Finish Predict..Sat Apr 4 08:50:10 2020 Time Total: 0: 00: 00.141096

[INFO] Start Predict..Sat Apr 4 08:50:10 2020 [INFO] Finish Predict..Sat Apr 4 08:50:10 2020 Time Total: 0: 00: 00.162114

[INFO] Start Predict..Sat Apr 4 08:50:10 2020 [INFO] Finish Predict..Sat Apr 4 08:50:10 2020 Time Total: 0: 00: 00.144101

[INFO] Start Predict..Sat Apr 4 08:50:10 2020 [INFO] Finish Predict..Sat Apr 4 08:50:10 2020 Time Total: 0: 00: 00.157125

I would be grateful for a hint :-)

Best regards Anja