easy_vision.python.custom_code

easy_vision.python.custom_code.custom_input

class easy_vision.python.custom_code.custom_input.CustomInput(dataset_config)

Bases: easy_vision.python.input.cv_input.CVInput

__init__(dataset_config)

Init input object

Parameters:dataset_config – easy-vision protobuf object, easy_vision.python.protos.dataset_pb2.DatasetConfig

classmethod create_class(name)

declare_feature_label()

decode_record(record_string)

decode data to tensor_dict

Parameters:record – record_data to be decoded Returns:a dict of tensors containing both features and labels Return type:tensor_dict

export_fn(export_config)

export_fn1(export_config)

easy_vision.python.custom_code.custom_model

class easy_vision.python.custom_code.custom_model.CustomClassificationModel(model_config, feature_dict, label_dict=None, mode='predict')

Bases: easy_vision.python.model.cv_model.CVModel

__init__(model_config, feature_dict, label_dict=None, mode='predict')

Init classification model

Parameters:

• model_config – a protobuf config object
• feature_dict – a dict of tensor, used as features
• label_dict – a dict of tensor, used as labels
• mode – indicates to build a graph for train, evaluate or predict
• categories – a list of dicts, each dist has two keys id and name

build_loss_graph()

calculate classification loss

Parameters:

• NxC tensor, where N is batch num, C is number of classes (self._prediction_dict['logits']) –
• N tensor (self._label_dict['groundtruth_image_classes']) –

Returns:

a dict of tensor containing different kinds of loss

Return type:

loss_dict

build_metric_graph(eval_config)

add metrics ops to graph

Parameters:eval_config – protobufer object, see python/protos/eval.proto. Returns:a dict of metric_op, each metric_op is a tuple of (update_op, value_op) Return type:metric_dict

build_predict_graph()

build classification network

Returns:a dict of prediction tensors

classmethod create_class(name)

get_outputs()

return a list of output key, which can be used to index output tensor in prediction_dict

class easy_vision.python.custom_code.custom_model.PyEvaluator

Bases: easy_vision.python.evaluation.evaluator.Evaluator

__init__()

Construct eval ops from tensor

Parameters:list of string, metric names this evaluator will return (metric_names) –

add_batch_image_info(predictions, labels)

store prediction and labels in a internal list

Parameters:

• batched prediction result, numpy array with shape N (predictions) –
• batched labels, numpy array with shape N (labels) –

clear()

clear internal storage

evaluate()

python evaluation code which will be run after all test batched data are predicted