Log model checkpoint as W&B Artifact. Args: model_path (str): Path of the checkpoint to log. aliases (list): List of

Log the W&B Tables for validation data as artifact and calls `use_artifact` on it so that the evaluation table can use the reference o

Log the W&B Tables for model evaluation. The table will be logged multiple times creating new version. Use this to compare models at

Build multiple branch.

Build fuse layer.

Make each layer.

Build one branch.

Make stem layer for ResNet.

Positioning embeding method. Resize the pos_embed, if the input image size doesn't match the training size. Args:

Placeholder for augmentation test.

Build DistributedDataParallel module by device type. If device is cuda, return a MMDistributedDataParallel module; if device is mlu, return a

build DataParallel module by device type. if device is cuda, return a MMDataParallel module; if device is mlu, return a MLUDataParallel modul

Build pixel sampler for segmentation map.

Build segmentor.

CAM feature classification.

Collect the information of the running environments.

Placeholder for extract features from images.

calculate the f-score value. Args: precision (float | torch.Tensor): The precision value. recall (float | torch.Tensor): The reca

This function is for finding the latest checkpoint. It will be used when automatically resume, modified from https://github.com/open-mmlab/mm

Rescale the initialization according to layer id. This function is copied from https://github.com/microsoft/unilm/blob/master/beit/modeling_

Placeholder of forward function.

Forward function.

Forward function.

Forward function.

generate 2d decay mask, the result is (HW)*(HW)

generate 2d decay mask, the result is (HW)*(HW) H, W are the numbers of patches at each column and row

generate 2d decay mask, the result is (HW)*(HW)

generate 2d decay mask, the result is (HW)*(HW) H, W are the numbers of patches at each column and row

Get the layer id to set the different learning rates in ``layer_wise`` decay_type. Args: var_name (str): The key of the model.

Get the layer id to set the different learning rates. Args: var_name (str): The key of the model. num_max_layer (int): Maximum nu

Sample points for testing. Find ``num_points`` most uncertain points from ``uncertainty_map``. Args: seg_logits (Tensor)

Sample points for training. Sample points in [0, 1] x [0, 1] coordinate space based on their uncertainty. The uncertainties are calcu

Get the stage id to set the different learning rates in ``stage_wise`` decay_type. Args: var_name (str): The key of the model.

Initialize weight of later layer.

Calculate intersection and Union. Args: pred_label (ndarray | str): Prediction segmentation map or predict result filename.

Returns a bool indicating if MLU is currently available.

Compute segmentation loss.

Compute segmentation loss.

Compute ``pam_cam``, ``pam``, ``cam`` loss.

Compute segmentation loss.

Stack InvertedResidual blocks to build a layer for MobileNetV2. Args: out_channels (int): out_channels of block. num_

Pack all blocks in a stage into a ``ResLayer``.

make plugins for ResNet 'stage_idx'th stage . Currently we support to insert 'context_block', 'empirical_attention_block', 'nonlocal_

Test model with multiple gpus by progressive mode. This method tests model with multiple gpus and collects the results under two different mo

nn.Module: the normalization layer named "norm0"

nn.Module: the normalization layer named "norm1"

nn.Module: the normalization layer named "norm1"

nn.Module: normalization layer after the first convolution layer

nn.Module: the normalization layer named "norm1"

nn.Module: normalization layer after the second convolution layer

nn.Module: normalization layer after the second convolution layer

PAM feature classification.

Forward function of PSP module.

Forward function of PSP module.

Reduce loss as specified. Args: loss (Tensor): Elementwise loss tensor. reduction (str): Options are "none", "mean" and "sum".

Reduce loss as specified. Args: loss (Tensor): Elementwise loss tensor. reduction (str): Options are "none", "mean" and "sum".

Reduce mean when distributed training.