run_accuracy_check.py

import argparse
import pickle
import numpy as np

from models import TensorflowModel, MovidiusModel, MovidiusModelV2, TPUModel
from utils import compute_tf_stats 

parser = argparse.ArgumentParser()
parser.add_argument("--cpu", action="store_true")
parser.add_argument("--gpu", action="store_true")
parser.add_argument("--tpu", action="store_true")
parser.add_argument("--movidius", action="store_true")
parser.add_argument("--movidius_2", action="store_true")
parser.add_argument("--mov_graph", type=str)
args = parser.parse_args()

# load parameters and data
with open('./data/test_data.pkl', 'rb') as pfile:
    test_data = pickle.load(pfile)

with open('./data/train_data.pkl', 'rb') as pfile:
    train_data = pickle.load(pfile)

if args.cpu or args.gpu:
    with open('./data/inference_weights.pkl', 'rb') as pfile:
        weights = pickle.load(pfile)

    # build the model using weights from previously trained model
    model = TensorflowModel(n_inputs=390, n_layers=2)
    model.build(with_gpu=args.gpu, n_copies=1)
    model.start_session()
    model.set_weights(weights)

elif args.movidius:
    model = MovidiusModel()
    model.load_graph(args.mov_graph)

elif args.movidius_2:
    model = MovidiusModelV2()

elif args.tpu:
    model = TPUModel()

else:
    raise Exception('No hardware specified to run accuracy check on!')

# print whole-word spotting accuracy
print('Training Data Statistics:')
compute_tf_stats(model, train_data)
print('')
print('Testing Data Statistics:')
compute_tf_stats(model, test_data)