deeplabv3_mobilenet_v3_large

torchvision.models.segmentation.deeplabv3_mobilenet_v3_large(*, weights: Optional[DeepLabV3_MobileNet_V3_Large_Weights] = None, progress: bool = True, num_classes: Optional[int] = None, aux_loss: Optional[bool] = None, weights_backbone: Optional[MobileNet_V3_Large_Weights] = MobileNet_V3_Large_Weights.IMAGENET1K_V1, **kwargs: Any) → DeepLabV3

Constructs a DeepLabV3 model with a MobileNetV3-Large backbone.

Reference: Rethinking Atrous Convolution for Semantic Image Segmentation.

Parameters:

• weights (DeepLabV3_MobileNet_V3_Large_Weights, optional) – The pretrained weights to use. See DeepLabV3_MobileNet_V3_Large_Weights below for more details, and possible values. By default, no pre-trained weights are used.

• progress (bool, optional) – If True, displays a progress bar of the download to stderr. Default is True.

• num_classes (int, optional) – number of output classes of the model (including the background)

• aux_loss (bool, optional) – If True, it uses an auxiliary loss

• weights_backbone (MobileNet_V3_Large_Weights, optional) – The pretrained weights for the backbone

• **kwargs – unused

class torchvision.models.segmentation.DeepLabV3_MobileNet_V3_Large_Weights(value)

The model builder above accepts the following values as the weights parameter. DeepLabV3_MobileNet_V3_Large_Weights.DEFAULT is equivalent to DeepLabV3_MobileNet_V3_Large_Weights.COCO_WITH_VOC_LABELS_V1. You can also use strings, e.g. weights='DEFAULT' or weights='COCO_WITH_VOC_LABELS_V1'.

DeepLabV3_MobileNet_V3_Large_Weights.COCO_WITH_VOC_LABELS_V1:

These weights were trained on a subset of COCO, using only the 20 categories that are present in the Pascal VOC dataset. Also available as DeepLabV3_MobileNet_V3_Large_Weights.DEFAULT.

miou (on COCO-val2017-VOC-labels)	60.3
pixel_acc (on COCO-val2017-VOC-labels)	91.2
categories	__background__, aeroplane, bicycle, … (18 omitted)
min_size	height=1, width=1
num_params	11029328
recipe	link
GFLOPS	10.45
File size	42.3 MB

The inference transforms are available at DeepLabV3_MobileNet_V3_Large_Weights.COCO_WITH_VOC_LABELS_V1.transforms and perform the following preprocessing operations: Accepts PIL.Image, batched (B, C, H, W) and single (C, H, W) image torch.Tensor objects. The images are resized to resize_size=[520] using interpolation=InterpolationMode.BILINEAR. Finally the values are first rescaled to [0.0, 1.0] and then normalized using mean=[0.485, 0.456, 0.406] and std=[0.229, 0.224, 0.225].