Shortcuts

vit_l_16

torchvision.models.vit_l_16(*, weights: Optional[ViT_L_16_Weights] = None, progress: bool = True, **kwargs: Any) VisionTransformer[source]

Constructs a vit_l_16 architecture from An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale.

Parameters:
  • weights (ViT_L_16_Weights, optional) – The pretrained weights to use. See ViT_L_16_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.vision_transformer.VisionTransformer base class. Please refer to the source code for more details about this class.

class torchvision.models.ViT_L_16_Weights(value)[source]

The model builder above accepts the following values as the weights parameter. ViT_L_16_Weights.DEFAULT is equivalent to ViT_L_16_Weights.IMAGENET1K_V1. You can also use strings, e.g. weights='DEFAULT' or weights='IMAGENET1K_V1'.

ViT_L_16_Weights.IMAGENET1K_V1:

These weights were trained from scratch by using a modified version of TorchVision’s new training recipe. Also available as ViT_L_16_Weights.DEFAULT.

acc@1 (on ImageNet-1K)

79.662

acc@5 (on ImageNet-1K)

94.638

categories

tench, goldfish, great white shark, … (997 omitted)

num_params

304326632

min_size

height=224, width=224

recipe

link

GFLOPS

61.55

File size

1161.0 MB

The inference transforms are available at ViT_L_16_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=[242] using interpolation=InterpolationMode.BILINEAR, 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].

ViT_L_16_Weights.IMAGENET1K_SWAG_E2E_V1:

These weights are learnt via transfer learning by end-to-end fine-tuning the original SWAG weights on ImageNet-1K data.

acc@1 (on ImageNet-1K)

88.064

acc@5 (on ImageNet-1K)

98.512

categories

tench, goldfish, great white shark, … (997 omitted)

recipe

link

license

link

num_params

305174504

min_size

height=512, width=512

GFLOPS

361.99

File size

1164.3 MB

The inference transforms are available at ViT_L_16_Weights.IMAGENET1K_SWAG_E2E_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=[512] using interpolation=InterpolationMode.BICUBIC, followed by a central crop of crop_size=[512]. 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].

ViT_L_16_Weights.IMAGENET1K_SWAG_LINEAR_V1:

These weights are composed of the original frozen SWAG trunk weights and a linear classifier learnt on top of them trained on ImageNet-1K data.

acc@1 (on ImageNet-1K)

85.146

acc@5 (on ImageNet-1K)

97.422

categories

tench, goldfish, great white shark, … (997 omitted)

recipe

link

license

link

num_params

304326632

min_size

height=224, width=224

GFLOPS

61.55

File size

1161.0 MB

The inference transforms are available at ViT_L_16_Weights.IMAGENET1K_SWAG_LINEAR_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].

Docs

Access comprehensive developer documentation for PyTorch

View Docs

Tutorials

Get in-depth tutorials for beginners and advanced developers

View Tutorials

Resources

Find development resources and get your questions answered

View Resources