Functions — github.com/Wuziyi616/LEOD

CodeHub MCPcopy Create free account

hub / github.com/Wuziyi616/LEOD / functions

Functions938 in github.com/Wuziyi616/LEOD

⨍ Functions 938◇ Types & classes 194↳ Endpoints 1

↓ 3 callersMethodupdate

Params: frame_idx - int dets - a numpy array of detections in shape (N, 4/5), each in the format of [x,y,w,h(

modules/tracking/linear.py:213

↓ 3 callersFunctionxywh2xyxy

Takes a bounding box in the form [x,y,w,h] and returns it in the form [x1,y1,x2,y2] where x1,y1 is the top left and x2,y2 is the bottom rig

modules/tracking/utils.py:62

↓ 2 callersMethod__init__

(self, mdl_config: DictConfig)

models/detection/recurrent_backbone/maxvit_rnn.py:26

↓ 2 callersMethod__init__

(self, in_channels, use_scale=True, rd_ratio=1/8, rd_channels=None, rd_divisor=8, **kwargs)

models/layers/maxvit/layers/non_local_attn.py:23

↓ 2 callersMethod__init__

(self, embed_len_decoder: int)

models/layers/maxvit/layers/ml_decoder.py:93

↓ 2 callersMethod__init__

(self, block_size=4)

models/layers/maxvit/layers/space_to_depth.py:6

↓ 2 callersMethod_add_to_buffer

(self, key: str, value: List[np.ndarray])

utils/evaluation/prophesee/evaluator.py:47

↓ 2 callersFunction_conservative_bbox_filter

Filter bbox that are too small by its height, width, or diag length.

modules/utils/ssod.py:90

↓ 2 callersFunction_create_fc

(num_features, num_classes, use_conv=False)

models/layers/maxvit/layers/classifier.py:22

↓ 2 callersFunction_create_pool

(num_features, num_classes, pool_type='avg', use_conv=False)

models/layers/maxvit/layers/classifier.py:11

↓ 2 callersFunction_draw_pred_bbox

(img, pred, color=(0, 255, 0))

vis_pred.py:131

↓ 2 callersMethod_get_bbox_loss_weight

Use predicted scores to weigh the bbox loss.

models/detection/yolox/models/yolo_head.py:359

↓ 2 callersMethod_get_from_buffer

(self, key: str)

utils/evaluation/prophesee/evaluator.py:55

↓ 2 callersMethod_has_states

modules/utils/detection.py:104

↓ 2 callersMethod_hflip_bbox

Apply horizontal flip to bboxes.

modules/utils/tta.py:130

↓ 2 callersMethod_load_range_labels

Load labels of frames in [start_idx, end_idx).

data/genx_utils/sequence_base.py:147

↓ 2 callersMethod_pad_tensor_impl

(input_tensor: th.Tensor, desired_hw: Tuple[int, int], mode: str, value: Any)

utils/padding.py:32

↓ 2 callersMethod_rand_another

2 cases: - not skipping labels: only happens when `time_flip` is True simply choose another frame (except the last one) - sk

data/genx_utils/sequence_rnd.py:123

↓ 2 callersMethod_reset_buffer

utils/evaluation/prophesee/evaluator.py:40

↓ 2 callersMethod_reset_buffer

callbacks/viz_base.py:30

↓ 2 callersFunction_split_channels

(num_chan, num_groups)

models/layers/maxvit/layers/mixed_conv2d.py:14

↓ 2 callersMethod_track

We track the bbox and filter out those from short tracklets.

modules/pseudo_labeler.py:202

↓ 2 callersFunction_trunc_normal_

(tensor, mean, std, a, b)

models/layers/maxvit/layers/weight_init.py:8

↓ 2 callersMethod_update_tflip

Update labels that undergo time-flip TTA.

modules/utils/tta.py:155

↓ 2 callersMethod_val_test_step_impl

(self, batch: Any, mode: Mode)

modules/detection.py:300

↓ 2 callersMethod_zero_zoom_in_state

Set zoom_in state as zeros.

data/utils/augmentor.py:277

↓ 2 callersFunctionadaptive_catavgmax_pool2d

(x, output_size=1)

models/layers/maxvit/layers/adaptive_avgmax_pool.py:30

↓ 2 callersMethodadd_ev_repr

Append `ev_repr[selected_indices]` to `self.repr_list`.

modules/utils/detection.py:74

↓ 2 callersMethodadd_to_buffer

(self, key: Enum, value: Union[np.ndarray, th.Tensor])

callbacks/viz_base.py:35

↓ 2 callersFunctionbuild_fourier_pos_embed

( feat_shape: List[int], bands: Optional[torch.Tensor] = None, num_bands: int = 64,

models/layers/maxvit/layers/pos_embed.py:72

↓ 2 callersFunctionbuild_streaming_evaluation_dataset

( datapipes: List[MapDataPipe], batch_size: int )

data/genx_utils/dataset_streaming.py:219

↓ 2 callersMethodclamp_to_frame_

data/genx_utils/labels.py:52

↓ 2 callersFunctioncollate

r""" General collate function that handles collection type of element within each batch and opens function registry to deal with speci

data/genx_utils/collate_from_pytorch.py:18

↓ 2 callersMethodcollect_data

modules/pseudo_labeler.py:497

↓ 2 callersFunctionconv2d_same

( x, weight: torch.Tensor, bias: Optional[torch.Tensor] = None, stride: Tuple[int, int] = (1, 1),

models/layers/maxvit/layers/conv2d_same.py:13

↓ 2 callersFunctioncreate_conv2d_pad

(in_chs, out_chs, kernel_size, **kwargs)

models/layers/maxvit/layers/conv2d_same.py:33

↓ 2 callersFunctioncustom_collate

data/genx_utils/collate.py:56

↓ 2 callersFunctioncv2_draw_bboxes

draw bboxes in the image img

utils/evaluation/prophesee/visualize/vis_utils.py:136

↓ 2 callersFunctiondetect_ckpt

Automatically detect checkpoints in the ckpt_path. Useful in SLURM preemption systems.

↓ 2 callersMethoddtype

data/utils/representations.py:30

↓ 2 callersMethodev_repr_to_img

(x: np.ndarray)

callbacks/viz_base.py:164

↓ 2 callersFunctioneval_one_dataset

(config: DictConfig)

↓ 2 callersFunctionevaluate_label

Evaluate the quality of filtered pseudo labels.

modules/utils/ssod.py:231

↓ 2 callersFunctionevaluate_list

(result_boxes_list, gt_boxes_list, height: int, width: i

utils/evaluation/prophesee/evaluation.py:5

↓ 2 callersMethodfinish

Delete all remaining trackers.

modules/tracking/tracker.py:33

↓ 2 callersMethodflip_lr

Apply horizontal flip to one or a list of data.

data/utils/ssod_augmentor.py:127

↓ 2 callersMethodget_assignments_w_ignore

( self, num_gt, gt_bboxes_per_image, # [n, 4] gt_classes, # [n] bbox

models/detection/yolox/models/yolo_head.py:975

↓ 2 callersFunctionget_condconv_initializer

(initializer, num_experts, expert_shape)

models/layers/maxvit/layers/cond_conv2d.py:21

↓ 2 callersFunctionget_dataloading_hw

(dataset_config: DictConfig)

data/utils/spatial.py:20

↓ 2 callersMethodget_embed

(self, shape: List[int])

models/layers/maxvit/layers/pos_embed.py:201

↓ 2 callersMethodget_ev_repr_as_list

(self, start_idx: int = 0, end_idx: Optional[int] = No

modules/utils/detection.py:84

↓ 2 callersFunctionget_event_representation_dir

(path: Path, ev_representation_name: str)

data/genx_utils/sequence_base.py:16

↓ 2 callersMethodget_from_buffer

(self, key: Enum)

callbacks/viz_base.py:45

↓ 2 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

models/detection/yolox/models/yolo_head.py:702

↓ 2 callersMethodget_l1_target

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

models/detection/yolox/models/yolo_head.py:599

↓ 2 callersFunctionget_label

(labels, objframe_idx_2_label_idx, objframe_idx, hw, ds_by2)

↓ 2 callersFunctionget_labels_npz_fn

Get labels npz file name.

data/utils/misc.py:11

↓ 2 callersFunctionget_norm_layer

models/layers/maxvit/layers/create_norm.py:32

↓ 2 callersFunctionget_objframe_idx_2_repr_idx

(path: Path, ev_representation_name: str)

data/genx_utils/sequence_base.py:22

↓ 2 callersFunctionget_original_hw

(dataset_type: DatasetType)

data/utils/spatial.py:16

↓ 2 callersFunctionget_same_padding

(x: int, k: int, s: int, d: int)

models/layers/maxvit/layers/padding.py:18

↓ 2 callersMethodget_xywh

(self, format_='center', add_class_id: bool = False)

data/genx_utils/labels.py:521

↓ 2 callersMethodget_xyxy

(self, add_class_id: bool = False)

data/genx_utils/labels.py:533

↓ 2 callersFunctiongroup_rms

(x, groups: int = 32, eps: float = 1e-5)

models/layers/maxvit/layers/evo_norm.py:90

↓ 2 callersFunctioninplace_abn

(x, weight, bias, running_mean, running_var, training=True, momentum=0.1, eps=1e-05, activ

models/layers/maxvit/layers/inplace_abn.py:10

↓ 2 callersFunctioninstance_rms

(x, eps: float = 1e-5)

models/layers/maxvit/layers/evo_norm.py:47

↓ 2 callersFunctioninv_freq_bands

( num_bands: int, temperature: float = 100000., step: int = 2, dtype: torch.dt

models/layers/maxvit/layers/pos_embed.py:22

↓ 2 callersFunctioninv_instance_rms

(x, eps: float = 1e-5)

models/layers/maxvit/layers/filter_response_norm.py:14

↓ 2 callersMethodis_empty

modules/utils/detection.py:57

↓ 2 callersFunctionis_scriptable

models/layers/maxvit/layers/config.py:63

↓ 2 callersMethodload_everything

data/genx_utils/dataset_rnd.py:89

↓ 2 callersMethodmerge_channel_and_bins

(self, representation: th.Tensor)

data/utils/representations.py:71

↓ 2 callersFunctionnorm_cdf

models/layers/maxvit/layers/weight_init.py:11

↓ 2 callersMethodonly_load_labels

data/genx_utils/dataset_rnd.py:86

↓ 2 callersFunctionparse_header

Parses the header of a dat file Args: - f file handle to a dat file return : - int position of the file cursor after the

utils/evaluation/prophesee/io/dat_events_tools.py:120

↓ 2 callersFunctionpartialclass

(cls, *args, **kwargs)

data/genx_utils/dataset_streaming.py:192

↓ 2 callersFunctionpred_one_seq

Run the model on one event sequence and visualize it.

vis_pred.py:179

↓ 2 callersMethodpredict

Advances the state vector and returns the predicted bbox estimate.

modules/tracking/linear.py:67

↓ 2 callersMethodpredict_dataloader

modules/data/genx.py:225

↓ 2 callersMethodrandomize_augmentation

Calls `randomize_augmentation()` for both strong and weak aug.

data/utils/ssod_augmentor.py:58

↓ 2 callersFunctionrel_logits_1d

Compute relative logits along one dimension As per: https://gist.github.com/aravindsrinivas/56359b79f0ce4449bcb04ab4b56a57a2 Originally from

models/layers/maxvit/layers/bottleneck_attn.py:28

↓ 2 callersFunctionrel_logits_1d

Compute relative logits along one dimension As per: https://gist.github.com/aravindsrinivas/56359b79f0ce4449bcb04ab4b56a57a2 Originally from

models/layers/maxvit/layers/halo_attn.py:30

↓ 2 callersMethodresize_mat

(self, x, t: int)

models/layers/maxvit/layers/non_local_attn.py:85

↓ 2 callersMethodreverse_flip_lr_

Reverse operation of flip_lr_.

data/genx_utils/labels.py:511

↓ 2 callersMethodreverse_zoom_in_and_rescale_

Reverse operation of zoom_in_and_rescale_. 1) Rescales the original canvas to a smaller canvas by a factor of 1/zoom_in_factor (down

data/genx_utils/labels.py:410

↓ 2 callersMethodreverse_zoom_out_and_rescale_

Reverse operation of zoom_out_and_rescale_. 1) Places the downscaled canvas back into the original position within the original canv

data/genx_utils/labels.py:459

↓ 2 callersMethodsimota_matching

(self, cost, pair_wise_ious, gt_classes, num_gt, fg_mask)

models/detection/yolox/models/yolo_head.py:734

↓ 2 callersMethodtest_dataloader

modules/data/genx.py:216

↓ 2 callersMethodtime_flip_

data/genx_utils/labels.py:707

↓ 2 callersMethodtime_flip_data

Flip the order of the sequence.

data/genx_utils/sequence_base.py:208

↓ 2 callersMethodupdate

Append new model predictions to self.

modules/pseudo_labeler.py:107

↓ 2 callersFunctionxyxy2xywh

Convert bounding box from xyxy to xywh format.

utils/bbox.py:75

↓ 2 callersMethodyield_pyramid_indices

E.g. input start_idx=0, end_idx=4, we get 0, 1, 2, 3, 3, 2, 1, 0, 0, 1, 2, 3, ...

data/utils/stream_sharded_datapipe.py:32

↓ 1 callersMethod__getitem__

(self, index: int)

data/genx_utils/dataset_rnd.py:95

↓ 1 callersMethod__init__

(self, path: str, scale_ratio: int, filter_config: DictConfig, postproc_cfg: DictConfig)

modules/pseudo_labeler.py:97

↓ 1 callersMethod__init__

(self, source_dp: IterDataPipe)

data/utils/stream_concat_datapipe.py:16

↓ 1 callersMethod__init__

(self, path: Path, ev_representation_name: str, sequence_le

data/genx_utils/sequence_streaming.py:62

↓ 1 callersMethod__init__

(self, object_labels: th.Tensor, input_size_hw: Tuple[int, int])

data/genx_utils/labels.py:252

↓ 1 callersMethod__init__

(self, path: Path, dataset_mode: DatasetMode, dataset_confi

data/genx_utils/dataset_rnd.py:27

↓ 1 callersMethod__init__

(self, reduction="none", loss_type="iou")

models/detection/yolox/models/losses.py:13

← previous next →101–200 of 938, ranked by callers