def _load_net(self):
# Load feature extractor
feature_extractor = feature_extractors.StridedInflatedEfficientNet()
checkpoint = engine.load_weights(self.feature_extractor_weights)
feature_extractor.load_state_dict(checkpoint)
feature_extractor.eval()
# Load a logistic regression classifier
gesture_classifier = LogisticRegression(
num_in=feature_extractor.feature_dim, num_out=30)
checkpoint = engine.load_weights(self.gesture_classifier_weights)
gesture_classifier.load_state_dict(checkpoint)
gesture_classifier.eval()
# Combine both models
self.net = Pipe(feature_extractor, gesture_classifier)
from realtimenet.downstream_tasks.gesture_recognition import INT2LAB
from realtimenet.downstream_tasks.nn_utils import Pipe, LogisticRegression
from realtimenet.downstream_tasks.postprocess import PostprocessClassificationOutput
if __name__ == "__main__":
# Parse arguments
args = docopt(__doc__)
camera_id = args['--camera_id'] or 0
path_in = args['--path_in'] or None
path_out = args['--path_out'] or None
title = args['--title'] or None
use_gpu = args['--use_gpu']
# Load feature extractor
feature_extractor = feature_extractors.StridedInflatedEfficientNet()
checkpoint = engine.load_weights('resources/backbone/strided_inflated_efficientnet.ckpt')
feature_extractor.load_state_dict(checkpoint)
feature_extractor.eval()
# Load a logistic regression classifier
gesture_classifier = LogisticRegression(num_in=feature_extractor.feature_dim,
num_out=30)
checkpoint = engine.load_weights('resources/gesture_detection/efficientnet_logistic_regression.ckpt')
gesture_classifier.load_state_dict(checkpoint)
gesture_classifier.eval()
# Concatenate feature extractor and met converter
net = Pipe(feature_extractor, gesture_classifier)
# Create inference engine, video streaming and display instances