"train_image_folder": "sample_datasets/segmentation/imgs",
"train_annot_folder": "sample_datasets/segmentation/anns",
"train_times": 4,
"valid_image_folder": "sample_datasets/segmentation/imgs_validation",
"valid_annot_folder": "sample_datasets/segmentation/anns_validation",
"valid_times": 4,
"valid_metric": "val_loss",
"batch_size": 8,
"learning_rate": 1e-4,
"saved_folder": "/home/ubuntu/space safety/segment",
"first_trainable_layer": "",
"ignore_zero_class": False,
"augumentation": True
},
"converter" : {
"type": ['k210']
}
}
dict = {'all':[classifier,detector,segnet],'classifier':[classifier],'detector':[detector],'segnet':[segnet]}
return dict[network_type]
for item in configs(args.type):
model_path = setup_training(config_dict=item)
K.clear_session()
setup_inference(item,model_path)
"learning_rate": 1e-4,
"saved_folder": "/home/ubuntu/space safety/segment",
"first_trainable_layer": "",
"ignore_zero_class": False,
"augumentation": True
},
"converter": {
"type": ["k210", "tflite"]
}
}
dict = {
'classifier': [classifier],
'detector': [detector],
'segnet': [segnet]
}
return dict[network_type]
#visualize_dataset('/home/ubuntu/github/sample_datasets/detector/imgs','/home/ubuntu/github/sample_datasets/detector/anns')
if not args.conf:
for item in configs(args.type):
setup_inference(item, args.weights)
K.clear_session()
else:
with open(args.conf) as config_buffer:
config = json.loads(config_buffer.read())
setup_inference(config, args.weights)