MCPcopy Create free account

hub / github.com/FudanCVL/MOVE / functions

Functions231 in github.com/FudanCVL/MOVE

⨍ Functions 231◇ Types & classes 62
Methodforward
Forward function. Args: x: input features with shape of (num_windows*B, N, C) mask: (0/-inf) mask with shape of (num_
libs/models/DMA/video_swin_referfomer/swin_transformer.py:127
Methodforward
Forward function. Args: x: Input feature, tensor size (B, H*W, C). H, W: Spatial resolution of the input feature.
libs/models/DMA/video_swin_referfomer/swin_transformer.py:202
Methodforward
Forward function. Args: x: Input feature, tensor size (B, H*W, C). H, W: Spatial resolution of the input feature.
libs/models/DMA/video_swin_referfomer/swin_transformer.py:273
Methodforward
Forward function. Args: x: Input feature, tensor size (B, H*W, C). H, W: Spatial resolution of the input feature.
libs/models/DMA/video_swin_referfomer/swin_transformer.py:362
Methodforward
Forward function.
libs/models/DMA/video_swin_referfomer/swin_transformer.py:427
Methodforward
Forward function.
libs/models/DMA/video_swin_referfomer/swin_transformer.py:595
Methodforward
(self, tensor_list: NestedTensor)
libs/models/DMA/video_swin_referfomer/swin_transformer.py:632
Methodforward
(self, tensor_list: NestedTensor)
libs/models/DMA/video_swin_referfomer/swin_transformer.py:665
FunctiongetPositionEncoding1d
(seq_length, feat_dim)
libs/models/DMA/DMA.py:1051
Methodget_consecutive_frames
Get consecutive frame sequence centered at center_frame Args: center_frame: center frame index num_frames: number of f
libs/dataset/MOVE.py:501
Methodget_metrics
获取所有评估指标
libs/utils/Logger.py:427
Functionget_model_para_number
(model)
libs/utils/Restore.py:57
Functionget_save_dir
(args)
libs/utils/Restore.py:64
Methodgetalltime
(self)
libs/utils/Logger.py:497
Methodgetloss
get every step loss and reset
libs/utils/Logger.py:302
Methodgettime
(self, epoch, begin_time)
libs/utils/Logger.py:475
Methodinflate_weights
Inflate the swin2d parameters to swin3d. The differences between swin3d and swin2d mainly lie in an extra axis. To utilize the pretra
libs/models/DMA/video_swin_referfomer/video_swin_transformer.py:561
Methodlogiou
(self, epoch=None, step=None)
libs/utils/Logger.py:421
Functionlr_lambda
(episode)
tools/train.py:80
Functionprocess_episode
Process a single episode
tools/evaluate.py:30
Methodrender
Show time remaining.
libs/utils/track_progress_rich.py:35
Functionrestore
(args, model, test_best=False)
libs/utils/Restore.py:10
Functionrestore_from_weight
(model,weight)
libs/utils/Restore.py:40
Functionsave_model
(args, epoch, model, optimizer, is_best=False)
libs/utils/Restore.py:79
Methodt1
(self)
libs/utils/Logger.py:485
Methodt2
(self)
libs/utils/Logger.py:488
Methodtrain
Convert the model into training mode while keep layers freezed.
libs/models/DMA/video_swin_referfomer/video_swin_transformer.py:629
Methodtraverse_hierarchy
(current_dict, current_path)
libs/dataset/MOVE.py:78
Methodupdate_evl
(self, idx, query_mask, pred)
libs/utils/Logger.py:360
Methodupdateloss
(self, loss_val, loss_part1=0, loss_part2=0, loss_part3=0, loss_part4=0)
libs/utils/Logger.py:291
Methodwith_pos_embed
(self, tensor, pos: Optional[Tensor])
libs/models/DMA/transformers.py:154
← previous201–231 of 231, ranked by callers