maxvit_t

torchvision.models.maxvit_t(*, weights: Optional[MaxVit_T_Weights] = None, progress: bool = True, **kwargs: Any) → MaxVit

Constructs a maxvit_t architecture from MaxViT: Multi-Axis Vision Transformer.

Parameters:

• weights (MaxVit_T_Weights, optional) – The pretrained weights to use. See MaxVit_T_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.

• **kwargs – parameters passed to the torchvision.models.maxvit.MaxVit base class. Please refer to the source code for more details about this class.

class torchvision.models.MaxVit_T_Weights(value)

The model builder above accepts the following values as the weights parameter. MaxVit_T_Weights.DEFAULT is equivalent to MaxVit_T_Weights.IMAGENET1K_V1. You can also use strings, e.g. weights='DEFAULT' or weights='IMAGENET1K_V1'.

MaxVit_T_Weights.IMAGENET1K_V1:

These weights reproduce closely the results of the paper using a similar training recipe. Also available as MaxVit_T_Weights.DEFAULT.

acc@1 (on ImageNet-1K)	83.7
acc@5 (on ImageNet-1K)	96.722
categories	tench, goldfish, great white shark, … (997 omitted)
num_params	30919624
min_size	height=224, width=224
recipe	link

The inference transforms are available at MaxVit_T_Weights.IMAGENET1K_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=[224] using interpolation=InterpolationMode.BICUBIC, followed by a central crop of crop_size=[224]. 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].