Turi Create
4.0
|
#include <ml/neural_net/compute_context.hpp>
#include <ml/neural_net/model_backend.hpp>
#include <ml/neural_net/model_spec.hpp>
#include <toolkits/object_detection/od_model_trainer.hpp>
Go to the source code of this file.
Namespaces | |
turi | |
SKD. | |
Functions | |
neural_net::image_augmenter::options | turi::object_detection::DarknetYOLOInferenceAugmentationOptions (int batch_size, int output_height, int output_width) |
neural_net::image_augmenter::options | turi::object_detection::DarknetYOLOTrainingAugmentationOptions (int batch_size, int output_height, int output_width) |
EncodedInputBatch | turi::object_detection::EncodeDarknetYOLO (InputBatch input_batch, size_t output_height, size_t output_width, size_t num_anchors, size_t num_classes) |
InferenceOutputBatch | turi::object_detection::DecodeDarknetYOLOInference (EncodedBatch batch, float confidence_threshold, float iou_threshold) |
Defines helper functions and the Model subclass for the darknet-yolo architecture.
Definition in file od_darknet_yolo_model_trainer.hpp.
neural_net::image_augmenter::options turi::object_detection::DarknetYOLOInferenceAugmentationOptions | ( | int | batch_size, |
int | output_height, | ||
int | output_width | ||
) |
Configures an image_augmenter for inference given darknet-yolo network parameters.
neural_net::image_augmenter::options turi::object_detection::DarknetYOLOTrainingAugmentationOptions | ( | int | batch_size, |
int | output_height, | ||
int | output_width | ||
) |
Configures an image_augmenter for training given darknet-yolo network parameters.
InferenceOutputBatch turi::object_detection::DecodeDarknetYOLOInference | ( | EncodedBatch | batch, |
float | confidence_threshold, | ||
float | iou_threshold | ||
) |
Decodes the raw inference output into structured predictions.
EncodedInputBatch turi::object_detection::EncodeDarknetYOLO | ( | InputBatch | input_batch, |
size_t | output_height, | ||
size_t | output_width, | ||
size_t | num_anchors, | ||
size_t | num_classes | ||
) |
Encodes the annotations of an input batch into the format expected by the darknet-yolo network.