Initializes counters for elapsed time and eta.

Returns param names which pass the filters specified in scheduler_cfg. Args: scheduler_cfg: The config for the scheduler paramete

Update memory with new frame data.

Directly turn binary `mask_inputs` into a output mask logits without using SAM. (same input and output shapes as in _forward_sam_head

Directly turn binary `mask_inputs` into a output mask logits without using SAM. (same input and output shapes as in _forward_sam_head

Directly turn binary `mask_inputs` into a output mask logits without using SAM. (same input and output shapes as in _forward_sam_head

Whether to use multimask output in the SAM head.

Whether to use multimask output in the SAM head.

Deprecated method. Please use `add_new_points_or_box` instead.

Add new points to a frame.

Add new points to a frame.

Wrapper over torch.distributed.all_reduce for performing min reduction of tensor over all processes in both distributed / non-distributed

Verifies that all the parameters matching the provided patterns are frozen - this acts as a safeguard when ignoring parameter when saving

Generates point grids for all crop layers.

Generates a 2D grid of points evenly spaced in [0,1]x[0,1].

Generates a 2D grid of points evenly spaced in [0,1]x[0,1].

Remove all input points or mask in a specific frame for a given object.

Remove all input points or mask in a specific frame for a given object.

Remove all input points or mask in a specific frame for a given object.

Remove all input points or mask in a specific frame for a given object.

This function recursively iterates through a dataset configuration, and collect all the dict_key that are defined

Constructs a VideoDatapoint sample to pass to transforms

Constructs a stochastic gradient descent or ADAM (or ADAMw) optimizer with momentum. i.e, constructs a torch.optim.Optimizer with zero-weight

Compute the DICE loss, similar to generalized IOU for masks Args: inputs: A float tensor of arbitrary shape. The pred

Remove from the state dictionary the parameters matching the provided unix patterns Args: patterns: the list of unix patterns to exc

Identical to the corresponding method in the parent (EfficientTAMVideoPredictor), but cloning the backbone features and pos encoding

Identical to the corresponding method in the parent (EfficientTAMVideoPredictor), but cloning the backbone features and pos encoding

Forward video tracking on each frame (and sample correction clicks).

Forward video tracking on each frame (and sample correction clicks).

Wrapper over torch.distributed.all_gather for performing 'gather' of 'tensor' over all processes in both distributed / non-distributed sc

Generates a list of crop boxes of different sizes. Each layer has (2**i)**2 boxes for the ith layer.

Get 1D sine positional embedding as in the original Transformer paper.

Get 1D sine positional embedding as in the original Transformer paper.

Calculate absolute positional embeddings. If needed, resize embeddings and remove cls_token dimension for the original embeddings. Ar

Calculate absolute positional embeddings. If needed, resize embeddings and remove cls_token dimension for the original embeddings. Ar

Return an activation function given a string

Return an activation function given a string

Produce a mapping from all the modules classes to the names of parames they own. Only counts a parameter as part of the immediate parent module,

Split a mask into per-object masks.

Args: inputs: A float tensor of arbitrary shape. The predictions for each example. targets: A float tensor with t

Filter masks at the edge of a crop, but not at the edge of the original image.

Loads a state dict into the given model. Args: state_dict: A dictionary containing the model's state dict, or a subset i

Load the video frames from video_path. The frames are resized to image_size as in the model and are loaded to GPU if offload_video_to_cpu=Fal

Load the video frames from video_path. The frames are resized to image_size as in the model and are loaded to GPU if offload_video_to_cpu=Fal

Load the video frames from a directory of JPEG files ("<frame_index>.jpg" format). The frames are resized to image_size x image_size and are

Load the video frames from a directory of JPEG files ("<frame_index>.jpg" format). The frames are resized to image_size x image_size and are

Load the video frames from a video file.

Load the video frames from a video file.