Swin Transformer#

torchgeo.models.swin_t(weights=None, *args, **kwargs)[source]#

Swin Transformer tiny model.

If you use this model in your research, please cite the following paper:

Added in version 0.8.

Parameters:

  • weights (Swin_T_Weights | None) – Pre-trained model weights to use.

  • *args (Any) – Additional arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

  • **kwargs (Any) – Additional keyword arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

Returns:

A Swin Transformer Tiny model.

Return type:

SwinTransformer

torchgeo.models.swin_s(weights=None, *args, **kwargs)[source]#

Swin Transformer small model.

If you use this model in your research, please cite the following paper:

Added in version 0.8.

Parameters:

  • weights (Swin_S_Weights | None) – Pre-trained model weights to use.

  • *args (Any) – Additional arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

  • **kwargs (Any) – Additional keyword arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

Returns:

A Swin Transformer Small model.

Return type:

SwinTransformer

torchgeo.models.swin_b(weights=None, *args, **kwargs)[source]#

Swin Transformer base model.

If you use this model in your research, please cite the following paper:

Added in version 0.8.

Parameters:

  • weights (Swin_B_Weights | None) – Pre-trained model weights to use.

  • *args (Any) – Additional arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

  • **kwargs (Any) – Additional keyword arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

Returns:

A Swin Transformer Base model.

Return type:

SwinTransformer

torchgeo.models.swin_v2_t(weights=None, *args, **kwargs)[source]#

Swin Transformer v2 tiny model.

If you use this model in your research, please cite the following paper:

Added in version 0.6.

Parameters:

  • weights (Swin_V2_T_Weights | None) – Pre-trained model weights to use.

  • *args (Any) – Additional arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

  • **kwargs (Any) – Additional keyword arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

Returns:

A Swin Transformer Tiny model.

Return type:

SwinTransformer

torchgeo.models.swin_v2_b(weights=None, *args, **kwargs)[source]#

Swin Transformer v2 base model.

If you use this model in your research, please cite the following paper:

Added in version 0.6.

Parameters:

  • weights (Swin_V2_B_Weights | None) – Pre-trained model weights to use.

  • *args (Any) – Additional arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

  • **kwargs (Any) – Additional keyword arguments to pass to torchvision.models.swin_transformer.SwinTransformer.

Returns:

A Swin Transformer Base model.

Return type:

SwinTransformer

class torchgeo.models.SwinBackbone_Weights(new_class_name, /, names, *, module=None, qualname=None, type=None, start=1, boundary=None)[source]#

Bases: WeightsEnum

SwinBackbone weights parent class.

These weights contain the encoder weights and optionally the backbone layernorm weights. To select whether layernorm weights are returned pass include_norms to get_state_dict (default is false).

Added in version 0.8.

get_state_dict(include_norms=False, *args, **kwargs)[source]#

Get the state dict for this model from provided url, optionally including backbone layernorm weights.

Parameters:

  • include_norms (bool) – Whether to also return backbone layernorm weights.

  • *args (Any) – anything passed to WeightsEnum get_state_dict.

  • **kwargs (Any) – anything passed to WeightsEnum get_state_dict.

Returns:

dict with state dict only if include_norms is False, dict with ‘state_dict’ and ‘feat_norms_state_dict’ if include_norms is True.

Return type:

Any

__new__(value)#
class torchgeo.models.Swin_T_Weights(*values)[source]#

Bases: SwinBackbone_Weights

Swin Transformer Tiny weights.

For torchvision swin_t implementation.

Added in version 0.8.

__new__(value)#
class torchgeo.models.Swin_S_Weights(*values)[source]#

Bases: SwinBackbone_Weights

Swin Transformer Small weights.

For torchvision swin_s implementation.

Added in version 0.8.

__new__(value)#
class torchgeo.models.Swin_B_Weights(*values)[source]#

Bases: SwinBackbone_Weights

Swin Transformer Base weights.

For torchvision swin_b implementation.

Added in version 0.8.

__new__(value)#
class torchgeo.models.Swin_V2_T_Weights(*values)[source]#

Bases: WeightsEnum

Swin Transformer v2 Tiny weights.

For torchvision swin_v2_t implementation.

Added in version 0.6.

__new__(value)#
class torchgeo.models.Swin_V2_B_Weights(*values)[source]#

Bases: WeightsEnum

Swin Transformer v2 Base weights.

For torchvision swin_v2_b implementation.

Added in version 0.6.

__new__(value)#