Create a list of default :class:`Augmentation` from config. Now it includes resizing and flipping. Returns: list[Augmentation]

Args: img_size (int): Input image size. patch_size (int): Patch size. in_chans (int): Number of input ima

Get relative positional embeddings according to the relative positions of query and key sizes. Args: q_size (int): size of qu

Add object_descriptions and logit (if applicable) to detectron2's instances_to_coco_json

Split a full s3 path into the bucket name and path.

Serializes this instance to a JSON string.

Add "gt_object_description" and "foreground" to detectron2's _match_and_label_boxes

Soft non-max suppression algorithm. Implementation of [Soft-NMS -- Improving Object Detection With One Line of Codec] (https://arxiv.org

Calculate decomposed Relative Positional Embeddings from :paper:`mvitv2`. https://github.com/facebookresearch/mvit/blob/19786631e330df9f3622e

Performs soft non-maximum suppression in a batched fashion. Each index value correspond to a category, and NMS will not be applied betwe

Given something that might be a URL (or might be a local path), determine which. If it's a URL, download the file and cache it, and retur

Add soft NMS to detectron2's fast_rcnn_inference_single_image

Constructs a `Config` from a Python dictionary of parameters.

Constructs a `BertConfig` from a json file of parameters.

r""" Instantiate a PretrainedConfig from a pre-trained model configuration. Params: **pretrained_model_name_or_path**: either:

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

Given a URL, look for the corresponding dataset in the local cache. If it's not there, download it. Then return the path to the cached file.

Calculate lr decay rate for different ViT blocks. Args: name (string): parameter name. lr_decay_rate (float): base lr decay r

Modified by Jialian Wu from https://github.com/facebookresearch/detectron2/blob/main/detectron2/engine/launch.py Launch multi-gpu or distribu

Launch multi-node training via deepspeed.

Load COCO class name text for object description for GRiT

Load Object365 class name text for object description for GRiT

Check ETag on S3 object.

Pull a file directly from S3.

Create configs and perform basic setups.

Create configs and perform basic setups.

Performs soft non-maximum suppression algorithm on axis aligned boxes Args: boxes (Tensor[N, 5]): boxes where NMS will be

Serializes this instance to a Python dictionary.

Save this instance to a json file.

Convert `url` into a hashed filename in a repeatable way. If `etag` is specified, append its hash to the url's, delimited by a period.

Partition into non-overlapping windows with padding if needed. Args: x (tensor): input tokens with [B, H, W, C]. window_size

Window unpartition into original sequences and removing padding. Args: x (tensor): input tokens with [B * num_windows, window_size, w

Args: dim (int): Number of input channels. num_heads (int): Number of attention heads. qkv_bias (bool: I

Args: in_channels (int): Number of input channels. out_channels (int): Number of output channels. bottlen

Args: dim (int): Number of input channels. num_heads (int): Number of attention heads in each ViT block.

Args: kernel_size (Tuple): kernel size of the projection layer. stride (Tuple): stride of the projection layer.