MCPcopy Create free account

hub / github.com/LiBingyu01/StitchFusion / functions

Functions314 in github.com/LiBingyu01/StitchFusion

⨍ Functions 314◇ Types & classes 106
↓ 16 callersMethod__init__
(self, c1, c2)
stitchfusion_with_tips_you_can_copy.py:10
↓ 16 callersMethod__init__
(self, c1, c2)
semseg/models/backbones/stitchfusion.py:10
↓ 9 callersFunctionget_train_augmentation
(size: Union[int, Tuple[int], List[int]], seg_fill: int = 0)
semseg/augmentations_mm.py:365
↓ 8 callersMethodupdate
(self, pred: Tensor, target: Tensor)
semseg/metrics.py:12
↓ 6 callersMethodcompute_f1
(self)
semseg/metrics.py:26
↓ 6 callersMethodcompute_iou
(self)
semseg/metrics.py:17
↓ 6 callersMethodcompute_pixel_acc
(self)
semseg/metrics.py:35
↓ 5 callersFunctionget_val_augmentation
(size: Union[int, Tuple[int], List[int]])
semseg/augmentations_mm.py:374
↓ 5 callersFunctionsetup_cudnn
()
semseg/utils/utils.py:26
↓ 4 callersMethod__init__
(self, optimizer, max_iter, decay_iter=1, power=0.9, last_epoch=-1)
semseg/schedulers.py:7
↓ 3 callersMethod__init__
delta: Controls weight given to FP and FN. This equals to dice score when delta=0.5
semseg/losses.py:48
↓ 3 callersMethod_open_img
(self, file)
semseg/datasets/DELIVER.py:103
↓ 3 callersMethod_open_img
(self, file)
semseg/datasets/kitti360.py:81
↓ 3 callersMethodget_lr
(self)
semseg/schedulers.py:13
↓ 2 callersMethodInfoNCE_loss
参数：pos: V*T-V*T 正样本矩阵 neg: V*T-V*T 负样本矩阵 dot: V*T-V*T 点乘结果矩阵 返回：loss 损失值
semseg/losses.py:338
↓ 2 callersMethod__init__
(self, dim, embed_dim)
semseg/models/heads/segformer.py:8
↓ 2 callersMethod_forward
(self, preds: Tensor, labels: Tensor)
semseg/losses.py:12
↓ 2 callersMethod_forward
(self, preds: Tensor, labels: Tensor)
semseg/losses.py:30
↓ 2 callersMethod_forward
(self, preds: Tensor, labels: Tensor)
semseg/losses.py:56
↓ 2 callersMethodcontrastive_loss
输入：feats T-C-V T: BATCH中的类/锚点数（可能重复） C: 特征空间的维度 V: 每个类/锚点的视图数，即每个类/锚点的样本数 labels T
semseg/losses.py:287
↓ 2 callersFunctionevaluate
(model, dataloader, device, loss_fn=None)
getting_start/val_mm.py:69
↓ 2 callersMethodget_masks
输入扁平化的标签，确定每个锚点的正负样本 参数：labels: T*V-1 num_anchors: T views_per_anchor: V 返回：pos_mask, neg_mask 正样
semseg/losses.py:320
↓ 2 callersFunctionhas_inf_or_nan
(x)
semseg/losses.py:88
↓ 2 callersFunctionnorm_cdf
(x)
semseg/models/layers/initialize.py:10
↓ 2 callersMethodsample_anchors_fast
input: dominant_classes N-1-H-W features N-C-H-W 返回采样的特征和标签： sampled_features T-C-V samp
semseg/losses.py:158
↓ 2 callersFunctiontime_sync
()
semseg/utils/utils.py:31
↓ 2 callersMethodtransform_val
(self, sample)
semseg/datasets/mcubes.py:133
↓ 1 callersMethod__init__
(self, p: float = None)
semseg/models/layers/common.py:23
↓ 1 callersMethod_get_file_names
(self, split_name)
semseg/datasets/nyu.py:82
↓ 1 callersMethod_get_file_names
(self, split_name)
semseg/datasets/mcubes.py:141
↓ 1 callersMethod_get_file_names
(self, split_name)
semseg/datasets/MFNet.py:73
↓ 1 callersMethod_get_file_names
(self, split_name)
semseg/datasets/kitti360.py:94
↓ 1 callersFunction_no_grad_trunc_normal_
(tensor, mean, std, a, b)
semseg/models/layers/initialize.py:7
↓ 1 callersMethod_open_img
(self, file)
semseg/datasets/nyu.py:70
↓ 1 callersMethod_open_img
(self, file)
semseg/datasets/MFNet.py:61
↓ 1 callersMethod_open_img
(self, file)
semseg/datasets/fmb.py:84
↓ 1 callersMethod_open_img
(self, file)
semseg/datasets/urbanlf.py:88
↓ 1 callersMethod_open_img
(self, file)
semseg/datasets/pst.py:73
↓ 1 callersMethod_select_views_per_class
(self, min_views, total_cls, cls_in_batch, cls_counts_in_batch)
semseg/losses.py:228
↓ 1 callersFunctioncal_flops
(model, modals, logger)
semseg/utils/utils.py:147
↓ 1 callersFunctioncleanup_ddp
()
semseg/utils/utils.py:78
↓ 1 callersMethodcm_contrastive_loss
输入: feats T-C-V T: 批次中的类别/锚点（可能重复） C: 特征空间的维度 V: 每个类别/锚点的视图数，即每个类别/锚点的样本数
semseg/losses.py:244
↓ 1 callersFunctiondraw_text
(image: torch.Tensor, seg_map: torch.Tensor, labels: list, fontsize: int = 15)
semseg/utils/visualize.py:77
↓ 1 callersMethodencode
(self, label: Tensor)
semseg/datasets/nyu.py:79
↓ 1 callersMethodencode
(self, label: Tensor)
semseg/datasets/MFNet.py:70
↓ 1 callersMethodencode
(self, label: Tensor)
semseg/datasets/fmb.py:93
↓ 1 callersMethodencode
(self, label: Tensor)
semseg/datasets/urbanlf.py:97
↓ 1 callersMethodencode
(self, label: Tensor)
semseg/datasets/pst.py:82
↓ 1 callersMethodencode
(self, label: Tensor)
semseg/datasets/DELIVER.py:112
↓ 1 callersMethodencode
(self, label: Tensor)
semseg/datasets/kitti360.py:90
↓ 1 callersFunctionevaluate
(model, dataloader, device, VIS_Saving, loss_fn=None)
getting_start/visulize_mm_label.py:72
↓ 1 callersFunctionevaluate
(model, dataloader, device, VIS_Saving, loss_fn=None)
getting_start/visulize_mm.py:72
↓ 1 callersFunctionevaluate
(model, dataloader, device, VIS_Saving, loss_fn=None)
getting_start/visulize_mm_tsne.py:75
↓ 1 callersFunctionevaluate_msf
(model, dataloader, device, scales, flip)
getting_start/val_mm.py:103
↓ 1 callersFunctionfix_seeds
(seed: int = 3407)
semseg/utils/utils.py:20
↓ 1 callersMethodget_dist_and_classes
确定每个scale*scale大小的地面真实标签块N-H-W中的类分布情况， 返回的类分布形状为N-C-H//scale-W//scale的tensor。 同时基于类分布确定每个地面真实标签块N-H-W中的【主导类】。 输出是N-
semseg/losses.py:140
↓ 1 callersFunctionget_logger
(log_file=None)
semseg/utils/utils.py:125
↓ 1 callersFunctionget_loss
(loss_fn_name: str = 'CrossEntropy', ignore_label: int = 255, cls_weights: Tensor = None)
semseg/losses.py:79
↓ 1 callersMethodget_lr_ratio
(self)
semseg/schedulers.py:32
↓ 1 callersMethodget_main_ratio
(self)
semseg/schedulers.py:35
↓ 1 callersFunctionget_optimizer
(model: nn.Module, optimizer: str, lr: float, weight_decay: float = 0.01)
semseg/optimizers.py:5
↓ 1 callersFunctionget_scheduler
(scheduler_name: str, optimizer, max_iter: int, power: int, warmup_iter: int, warmup_ratio: float)
semseg/schedulers.py:87
↓ 1 callersMethodget_warmup_ratio
(self)
semseg/schedulers.py:38
↓ 1 callersMethodinit_pretrained
(self, pretrained: str = None)
semseg/models/base.py:60
↓ 1 callersMethodinit_pretrained
(self, pretrained: str = None)
semseg/models/stitchfusion.py:19
↓ 1 callersFunctionload_dualpath_model
(model, model_file)
semseg/models/base.py:8
↓ 1 callersFunctionmain
(cfg)
getting_start/val_mm.py:138
↓ 1 callersFunctionmain
(cfg, save_dir)
getting_start/train_mm.py:32
↓ 1 callersFunctionmain
(cfg)
getting_start/visulize_mm_label.py:188
↓ 1 callersFunctionmain
(cfg)
getting_start/visulize_mm.py:161
↓ 1 callersFunctionmain
(cfg)
getting_start/visulize_mm_tsne.py:229
↓ 1 callersMethodmodel_forward
(self, img: Tensor)
getting_start/infer_mm.py:77
↓ 1 callersFunctionpad_image
(img, target_size)
getting_start/val_mm.py:28
↓ 1 callersFunctionpad_image
(img, target_size)
getting_start/visulize_mm_label.py:31
↓ 1 callersFunctionpad_image
(img, target_size)
getting_start/visulize_mm.py:31
↓ 1 callersFunctionpad_image
(img, target_size)
getting_start/visulize_mm_tsne.py:34
↓ 1 callersMethodpostprocess
(self, orig_img: Tensor, seg_map: Tensor, overlay: bool)
getting_start/infer_mm.py:61
↓ 1 callersMethodpredict
(self, img_fname: str, overlay: bool)
getting_start/infer_mm.py:80
↓ 1 callersMethodpreprocess
(self, image: Tensor)
getting_start/infer_mm.py:46
↓ 1 callersFunctionprint_iou
(epoch, iou, miou, acc, macc, class_names)
semseg/utils/utils.py:163
↓ 1 callersFunctionsliding_predict
(model, image, num_classes, flip=True)
getting_start/val_mm.py:35
↓ 1 callersFunctionsliding_predict
(model, image, num_classes, flip=True)
getting_start/visulize_mm.py:38
↓ 1 callersFunctionsliding_predict
(model, image, num_classes, flip=True)
getting_start/visulize_mm_tsne.py:41
↓ 1 callersMethodtransform_tr
(self, sample)
semseg/datasets/mcubes.py:123
↓ 1 callersFunctiontrunc_normal_
r"""Fills the input Tensor with values drawn from a truncated normal distribution. The values are effectively drawn from the normal distributi
semseg/models/layers/initialize.py:43
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:13
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:30
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:44
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:66
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:75
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:85
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:96
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:106
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:116
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:126
Method__call__
(self, image, label)
semseg/augmentations.py:133
Method__call__
(self, image, label)
semseg/augmentations.py:141
Method__call__
(self, img, label)
semseg/augmentations.py:153
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:174
Method__call__
(self, img: Tensor, mask: Tensor)
semseg/augmentations.py:191
next →1–100 of 314, ranked by callers