stages Namespace
Definition
Classes Index
| struct | BoxDecodeOptions |
|
Options driving a standalone BoxDecode invocation. More... | |
Functions Index
| TensorList | Tensors (const simaai::neat::Sample &sample) |
|
Extract the tensor list embedded in a single Sample. More... | |
| TensorList | Preproc (const std::vector< cv::Mat > &inputs, const simaai::neat::Model &model) |
|
Run only the preprocessing stage on raw cv::Mat inputs. More... | |
| TensorList | Infer (const TensorList &inputs, const simaai::neat::Model &model) |
|
Run only the inference stage on already-preprocessed tensors. More... | |
| TensorList | MLA (const TensorList &inputs, const simaai::neat::Model &model) |
| TensorList | Postprocess (const TensorList &inputs, const simaai::neat::Model &model) |
|
Run only the model's postprocessing stage on inference outputs. More... | |
| Sample | Preproc (const Sample &inputs, const simaai::neat::Model &model) |
|
Sample-list overload of the preprocessing stage. More... | |
| Sample | Infer (const Sample &inputs, const simaai::neat::Model &model) |
|
Sample-list overload of the inference stage. More... | |
| Sample | MLA (const Sample &inputs, const simaai::neat::Model &model) |
|
Sample-list overload of the MLA stage. More... | |
| Sample | Postprocess (const Sample &inputs, const simaai::neat::Model &model) |
|
Sample-list overload of the postprocessing stage. More... | |
| Sample | BoxDecode (const Sample &inputs, const simaai::neat::Model &model, const BoxDecodeOptions &opt) |
| BoxDecodeResultList | BoxDecodeResults (const Sample &inputs, const simaai::neat::Model &model, const BoxDecodeOptions &opt) |
|
Run BoxDecode and parse each output into typed bounding boxes. More... | |
Functions
BoxDecode()
|
Run the BoxDecode stage on inference outputs.
- Parameters
-
inputs Inference output samples to decode.
model Model providing anchors/strides/class metadata.
opt Decode-family selection and filtering thresholds.
- Returns
Sample list whose payloads carry decoded detections.
Definition at line 83 of file StageRun.h.
BoxDecodeResults()
|
Run BoxDecode and parse each output into typed bounding boxes.
This is the structured-result companion to BoxDecode(...). It preserves the list-only public API rule: pass Sample{sample} for a single inference output, and read results.front() when only one result is expected.
Definition at line 93 of file StageRun.h.
Infer()
|
Run only the inference stage on already-preprocessed tensors.
Definition at line 62 of file StageRun.h.
Infer()
|
Sample-list overload of the inference stage.
Definition at line 70 of file StageRun.h.
MLA()
|
Run only the MLA leg on already-prepared tensors.
Definition at line 64 of file StageRun.h.
MLA()
|
Sample-list overload of the MLA stage.
Definition at line 72 of file StageRun.h.
Postprocess()
|
Run only the model's postprocessing stage on inference outputs.
Definition at line 66 of file StageRun.h.
Postprocess()
|
Sample-list overload of the postprocessing stage.
Definition at line 74 of file StageRun.h.
Preproc()
|
Run only the preprocessing stage on raw cv::Mat inputs.
Definition at line 60 of file StageRun.h.
Preproc()
|
Sample-list overload of the preprocessing stage.
Definition at line 68 of file StageRun.h.
Tensors()
|
Extract the tensor list embedded in a single Sample.
Definition at line 58 of file StageRun.h.
The documentation for this namespace was generated from the following file:
Generated via doxygen2docusaurus 2.0.0 by Doxygen 1.9.8.