A simple benchmarks of TensorFlow lite model evaluation, focusing on the invoke step of the process.
To execute from a command prompt, run:
python evaluate.py
Example output:
c:\GitHub\Benchmark-TensorFlow-Windows>python evaluate.py
Python version: 3.7.4 (default, Aug 9 2019, 18:34:13) [MSC v.1915 64 bit (AMD64)]
2020-04-26 12:48:37: Initizliaing model
input_details=[{'name': 'normalized_input_image_tensor', 'index': 578, 'shape': array([ 1, 192, 192, 3]), 'dtype': <class 'numpy.uint8'>, 'quantization': (0.0078125, 128)}]
2020-04-26 12:48:37: Init done
Img 0, Eval time=356 ms. Invoke=352.
Img 1, Eval time=339 ms. Invoke=335.
Img 2, Eval time=338 ms. Invoke=335.
Img 3, Eval time=310 ms. Invoke=307.
Img 4, Eval time=315 ms. Invoke=311.
Img 5, Eval time=333 ms. Invoke=330.
Img 6, Eval time=303 ms. Invoke=301.
Img 7, Eval time=360 ms. Invoke=358.
Img 8, Eval time=293 ms. Invoke=290.
Img 9, Eval time=231 ms. Invoke=229.
Img 0, Eval time=244 ms. Invoke=242.
Img 1, Eval time=223 ms. Invoke=221.
Img 2, Eval time=220 ms. Invoke=218.