Functions — github.com/Nota-NetsPresso/netspresso-trainer

hub / github.com/Nota-NetsPresso/netspresso-trainer / functions

Functions1,065 in github.com/Nota-NetsPresso/netspresso-trainer

⨍ Functions 1,065◇ Types & classes 251

↓ 1 callersMethod_is_apply_noise

Return True if scheduler in noise range.

src/netspresso_trainer/schedulers/base.py:109

↓ 1 callersMethod_is_ready

(self)

src/netspresso_trainer/pipelines/train.py:97

↓ 1 callersMethod_is_ready

(self)

src/netspresso_trainer/pipelines/evaluation.py:64

↓ 1 callersMethod_is_ready

(self)

src/netspresso_trainer/pipelines/inference.py:56

↓ 1 callersMethod_make_block

(self, stage_param, in_channels)

src/netspresso_trainer/models/backbones/experimental/mobilevit.py:302

↓ 1 callersMethod_make_divisible

(value, divisor=8)

src/netspresso_trainer/models/backbones/experimental/mixnet.py:263

↓ 1 callersMethod_make_inverted_residual_blocks

(self, num_blocks, in_channels, out_channels, stride, expand_ratio)

src/netspresso_trainer/models/backbones/experimental/mobilevit.py:327

↓ 1 callersMethod_make_mobilevit_blocks

(self, num_transformer_blocks, in_channels, out_channels, stride, expand_ratio,

src/netspresso_trainer/models/backbones/experimental/mobilevit.py:341

↓ 1 callersMethod_make_stage

(self, in_channels: int, out_channels: int, num_layers: int)

src/netspresso_trainer/models/backbones/experimental/shufflenetv2.py:74

↓ 1 callersMethod_map_coordinates

Mapping keypoint coordinates into SimCC space.

src/netspresso_trainer/losses/pose_estimation/rtmcc.py:54

↓ 1 callersMethod_pad_1x1_to_3x3_tensor

(self, kernel1x1)

src/netspresso_trainer/models/op/custom.py:249

↓ 1 callersFunction_params_and_flops_fvcore

(model: nn.Module, sample_input: torch.Tensor)

src/netspresso_trainer/utils/stats.py:36

↓ 1 callersMethod_prepare_input

Prepare input tensor for inference. Args: x (Union[np.ndarray, torch.Tensor]): Input tensor Returns:

src/netspresso_trainer/models/base.py:260

↓ 1 callersMethod_process_list_items

(self, list_items, parent_key, sep=".")

src/netspresso_trainer/loggers/mlflow.py:222

↓ 1 callersMethod_process_output

Process the output of the interpreter.

src/netspresso_trainer/models/base.py:242

↓ 1 callersMethod_reset_parameters

(self)

src/netspresso_trainer/models/necks/experimental/rtdetr_hybrid_encoder.py:176

↓ 1 callersMethod_reset_parameters

(self)

src/netspresso_trainer/models/heads/detection/experimental/rtdetr_head.py:133

↓ 1 callersMethod_reset_parameters

(self)

src/netspresso_trainer/models/heads/detection/experimental/rtdetr_head.py:439

↓ 1 callersMethod_rotate_point

Rotate a point by an angle.

src/netspresso_trainer/dataloaders/augmentation/custom/image_proc.py:885

↓ 1 callersMethod_round_repeats

(repeats, depth_multi)

src/netspresso_trainer/models/backbones/experimental/mixnet.py:275

↓ 1 callersFunction_save_onnx

(model: nn.Module, f: Union[str, Path], sample_input: Tensor, opset_version=13, input_names='im

src/netspresso_trainer/utils/onnx.py:29

↓ 1 callersMethod_setup_mlflow

(self)

src/netspresso_trainer/loggers/mlflow.py:43

↓ 1 callersMethod_step_without_given_epoch

(self)

src/netspresso_trainer/schedulers/cosine_warm_restart.py:116

↓ 1 callersMethod_validate_params

(self, params: DictConfig)

src/netspresso_trainer/models/heads/detection/experimental/yolo_head.py:122

↓ 1 callersFunction_voc_color_map

(N=256, normalized=False)

tools/device_runtime/visualizers/detection_visualizer.py:21

↓ 1 callersFunction_yaml_for_logging

(config: DictConfig)

src/netspresso_trainer/utils/logger.py:89

↓ 1 callersFunctionaccuracy

Computes the precision@k for the specified values of k

src/netspresso_trainer/losses/detection/rtdetr.py:392

↓ 1 callersFunctionaccuracy_topk

Computes the accuracy over the k top predictions for the specified values of k

src/netspresso_trainer/metrics/classification/metric.py:26

↓ 1 callersFunctionadd_stream_handler

(level: str, distributed: bool)

src/netspresso_trainer/utils/logger.py:56

↓ 1 callersFunctionadjust_box_anns

(bbox, scale_ratio, padw, padh, w_max, h_max)

src/netspresso_trainer/dataloaders/augmentation/custom/mosaic.py:28

↓ 1 callersFunctionapply_affine_to_bboxes

(targets, target_size, M, scale)

src/netspresso_trainer/dataloaders/augmentation/custom/mosaic.py:103

↓ 1 callersFunctionas_grid

(image_batch, rows: Optional[int] = None, cols: Optional[int] = None)

src/netspresso_trainer/loggers/tensorboard.py:155

↓ 1 callersFunctionas_tuple

(tuple_string: str)

src/netspresso_trainer/dataloaders/utils/misc.py:91

↓ 1 callersFunctionbboxes_iou

(bboxes_a, bboxes_b, xyxy=True)

src/netspresso_trainer/utils/bbox_utils.py:106

↓ 1 callersFunctionbias_init_with_prob

initialize conv/fc bias value according to a given probability value.

src/netspresso_trainer/models/heads/detection/experimental/rtdetr_head.py:86

↓ 1 callersFunctionbox_cxcywh_to_xyxy

(x)

src/netspresso_trainer/models/op/denoising.py:28

↓ 1 callersFunctionbox_iou_batch

Compute Intersection over Union (IoU) of two sets of bounding boxes - `boxes_true` and `boxes_detection`. Both sets of boxes are

src/netspresso_trainer/metrics/detection/metric.py:31

↓ 1 callersFunctionbuild_ema

(model, conf)

src/netspresso_trainer/utils/model_ema.py:68

↓ 1 callersFunctionbuild_overwrited_dict

(optimizer_conf)

src/netspresso_trainer/optimizers/builder.py:26

↓ 1 callersFunctionbuild_postprocessor

(task: str, conf_model)

src/netspresso_trainer/postprocessors/builder.py:20

↓ 1 callersMethodcache_dataset

(self)

src/netspresso_trainer/dataloaders/base.py:102

↓ 1 callersFunctioncalculate_iou

(bbox1, bbox2, metric="iou", EPS=1e-7)

src/netspresso_trainer/utils/bbox_utils.py:132

↓ 1 callersMethodcalibrate

(self, pred, target, **kwargs)

src/netspresso_trainer/metrics/base.py:33

↓ 1 callersMethodchannel_shuffle

(self, x: torch.Tensor)

src/netspresso_trainer/models/op/custom.py:1024

↓ 1 callersMethodclear_training_targets

Clear internal training targets after training step

src/netspresso_trainer/models/heads/detection/experimental/rtdetr_head.py:667

↓ 1 callersMethodcompress

(self, model_name: str, model_path: Union[Path, str], batch_size: int, channels: int, height:

demo/func/pynetspresso.py:32

↓ 1 callersFunctioncompute_average_precision

Compute the average precision using 101-point interpolation (COCO), given the recall and precision curves. Args: recall (np.

src/netspresso_trainer/metrics/detection/metric.py:105

↓ 1 callersMethodcompute_loss

(self, out, target, matched_idxs)

src/netspresso_trainer/losses/detection/retinanet.py:42

↓ 1 callersFunctionconvert_graphmodule

(model: nn.Module)

src/netspresso_trainer/utils/fx.py:32

↓ 1 callersMethodconvert_to_deploy

(self)

src/netspresso_trainer/models/op/custom.py:228

↓ 1 callersMethodcriterion

Criterion function.

src/netspresso_trainer/losses/pose_estimation/rtmcc.py:118

↓ 1 callersFunctioncxcywh2cxcywhn

(cx, cy, w, h, img_w, img_h)

tools/open_dataset_tool/objects365.py:79

↓ 1 callersFunctiondataset_path_check

(conf_data: DictConfig, mode: Literal['train', 'test'])

src/netspresso_trainer/dataloaders/builder.py:35

↓ 1 callersMethoddeploy

(self, )

src/netspresso_trainer/models/base.py:80

↓ 1 callersMethoddevice

(self)

src/netspresso_trainer/models/full/experimental/pidnet.py:161

↓ 1 callersFunctiondownload_file

(url, path, max_retries=30, delay=5, silent=True)

tools/open_dataset_tool/objects365.py:82

↓ 1 callersFunctiondownload_model_checkpoint

( model_name: str, task: Optional[str] = None, # TODO: Pretrained weights can be distinguished by tas

src/netspresso_trainer/models/utils.py:200

↓ 1 callersMethoddraw

(self, image, detections, model_input_shape)

tools/device_runtime/visualizers/detection_visualizer.py:47

↓ 1 callersFunctiondrop_path

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks). Comment by Ross Wightman: This is the same as the D

src/netspresso_trainer/models/op/depth.py:22

↓ 1 callersMethodepoch_with_checkpoint_saving

(self, epoch: int)

src/netspresso_trainer/pipelines/train.py:112

↓ 1 callersMethodepoch_with_valid_logging

(self, epoch: int)

src/netspresso_trainer/pipelines/train.py:107

↓ 1 callersMethodevaluation

(self)

src/netspresso_trainer/pipelines/evaluation.py:73

↓ 1 callersFunctionevaluation_cli

()

src/netspresso_trainer/evaluator_main.py:104

↓ 1 callersFunctionevaluation_cli_without_additional_gpu_check

()

src/netspresso_trainer/evaluator_main.py:120

↓ 1 callersFunctionevaluation_with_yaml_impl

(gpus: Optional[Union[List, int]], data: Union[Path, str], augmentation: Union[Path, str],

src/netspresso_trainer/evaluator_main.py:68

↓ 1 callersMethodfilter_duplicates

Filter the maximum suitability target index of each anchor. Args: target_matrix [batch x targets x anchors]: The suitabi

src/netspresso_trainer/losses/detection/yolov9.py:178

↓ 1 callersMethodfilter_topk

Filter the top-k suitability of targets for each anchor. Args: target_matrix [batch x targets x anchors]: The suitabilit

src/netspresso_trainer/losses/detection/yolov9.py:160

↓ 1 callersMethodflatten_dict

(self, nested_dict, parent_key="", sep=".")

src/netspresso_trainer/loggers/mlflow.py:208

↓ 1 callersMethodfolding

(self, patches: Tensor, info_dict: Dict)

src/netspresso_trainer/models/backbones/experimental/mobilevit.py:167

↓ 1 callersMethodforward_ffn

(self, tgt)

src/netspresso_trainer/models/heads/detection/experimental/rtdetr_head.py:253

↓ 1 callersFunctionforward_model

(interpreter, input_data)

tools/device_runtime/tflite_run.py:54

↓ 1 callersMethodforward_others

( self, x, indices: Optional[Tensor] = None, *args, **kwargs )

src/netspresso_trainer/models/op/custom.py:688

↓ 1 callersMethodforward_patch_last

( self, x, indices: Optional[Tensor] = None, *args, **kwargs )

src/netspresso_trainer/models/op/custom.py:673

↓ 1 callersMethodgaussian_smoothng

(self, keypoints)

src/netspresso_trainer/losses/pose_estimation/rtmcc.py:68

↓ 1 callersMethodgenerate_anchors

( self, scales: List[int], aspect_ratios: List[float], dtype: torch.dtype = to

src/netspresso_trainer/models/heads/detection/experimental/detection/anchor_generator.py:77

↓ 1 callersFunctionget_affine_matrix

( target_size, degrees, translate, scales, shear, )

src/netspresso_trainer/dataloaders/augmentation/custom/mosaic.py:67

↓ 1 callersMethodget_cdn_matched_indices

get_cdn_matched_indices

src/netspresso_trainer/losses/detection/rtdetr.py:361

↓ 1 callersMethodget_cls_matrix

Get the (predicted class' probabilities) corresponding to the target classes across all anchors Args: predict_cls [batch

src/netspresso_trainer/losses/detection/yolov9.py:132

↓ 1 callersFunctionget_contrastive_denoising_training_group

cnd

src/netspresso_trainer/models/op/denoising.py:42

↓ 1 callersFunctionget_dataframe

(experiment_dir=OUTPUT_DIR)

demo/func/experiments.py:106

↓ 1 callersMethodget_decay

(self)

src/netspresso_trainer/utils/model_ema.py:54

↓ 1 callersMethodget_epoch_values

(self, epoch: int)

src/netspresso_trainer/schedulers/base.py:76

↓ 1 callersMethodget_equivalent_kernel_bias

(self)

src/netspresso_trainer/models/op/custom.py:238

↓ 1 callersMethodget_geometry_constraint

Calculate whether the center of an object is located in a fixed range of an anchor. This is used to avert inappropriate matching. It

src/netspresso_trainer/losses/detection/yolox.py:374

↓ 1 callersFunctionget_gpu_from_config

(conf_environment: DictConfig)

src/netspresso_trainer/utils/engine_utils.py:166

↓ 1 callersMethodget_iou_matrix

Get the IoU between each target bounding box and each predicted bounding box. Args: predict_bbox [batch x predicts x 4]:

src/netspresso_trainer/losses/detection/yolov9.py:148

↓ 1 callersMethodget_l1_target

(self, l1_target, gt, stride, x_shifts, y_shifts, eps=1e-8)

src/netspresso_trainer/losses/detection/yolox.py:109

↓ 1 callersMethodget_lr

(self)

src/netspresso_trainer/schedulers/cosine_warm_restart.py:90

↓ 1 callersFunctionget_model_config_path_list

(config_path_or_dir: Path)

tools/onnx_convert.py:52

↓ 1 callersFunctionget_model_config_path_list

(config_path_or_dir: Path)

tools/exir_convert.py:56

↓ 1 callersFunctionget_model_config_path_list

(config_path_or_dir: Path)

tools/fx_convert.py:44

↓ 1 callersFunctionget_mosaic_coordinate

(mosaic_image, mosaic_index, xc, yc, w, h, input_h, input_w)

src/netspresso_trainer/dataloaders/augmentation/custom/mosaic.py:34

↓ 1 callersMethodget_output

(self, output, anchor_grid, scaler)

src/netspresso_trainer/losses/detection/yolov9.py:333

↓ 1 callersMethodget_output_and_grid

(self, output, k, stride, dtype)

src/netspresso_trainer/losses/detection/yolo.py:39

↓ 1 callersMethodget_output_and_grid

(self, output, k, stride, dtype)

src/netspresso_trainer/losses/detection/yolox.py:439

↓ 1 callersMethodget_simcc_maximum

(self, simcc_x, simcc_y, apply_softmax: bool = False)

src/netspresso_trainer/postprocessors/pose_estimation.py:33

↓ 1 callersFunctionget_snippet_from_url

(url: str)

demo/app.py:48

↓ 1 callersMethodget_stats

(self)

src/netspresso_trainer/dataloaders/detection.py:233

↓ 1 callersMethodget_update_values

(self, num_updates: int)

src/netspresso_trainer/schedulers/base.py:79

↓ 1 callersMethodget_valid_matrix

Get a boolean mask that indicates whether each target bounding box overlaps with each anchor. Args: target_bbox [batch x

src/netspresso_trainer/losses/detection/yolov9.py:115

← previous next →201–300 of 1,065, ranked by callers