glasses.models.classification.deit package

Module contents

class glasses.models.classification.deit.DeiT(*args, head: torch.nn.modules.module.Module = <class 'glasses.models.classification.deit.DeiTClassificationHead'>, tokens: torch.nn.modules.module.Module = <class 'glasses.models.classification.deit.DeiTTokens'>, **kwargs)

Bases: glasses.models.classification.vit.ViT

Implementation of DeiT proposed in Training data-efficient image transformers & distillation through attention

An attention based distillation is proposed where a new token is added to the model, the dist token.

DeiT.deit_tiny_patch16_224()
DeiT.deit_small_patch16_224()
DeiT.deit_base_patch16_224()
DeiT.deit_base_patch16_384()

Parameters

• head (nn.Module, optional) – [description]. Defaults to DeiTClassificationHead.

• tokens (nn.Module, optional) – [description]. Defaults to DeiTTokens.

Initializes internal Module state, shared by both nn.Module and ScriptModule.

classmethod deit_base_patch16_224(**kwargs)

classmethod deit_base_patch16_384(**kwargs)

classmethod deit_small_patch16_224(**kwargs)

classmethod deit_tiny_patch16_224(**kwargs)

class glasses.models.classification.deit.DeiTClassificationHead(emb_size: int = 768, n_classes: int = 1000)

Bases: torch.nn.modules.module.Module

DeiT classification head, it relies on two heads using the cls and the`dist` token respectively. At test time, the prediction is made by avering the results from the two, while during training both predictions are returned.

Parameters

• emb_size (int, optional) – Embedding dimensions Defaults to 768.

• n_classes (int, optional) – [description]. Defaults to 1000.

forward(x: torch.Tensor) → torch.Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

training: bool

class glasses.models.classification.deit.DeiTTokens(emb_size: int)

Bases: glasses.models.classification.vit.ViTTokens

Tokens for DeiT, it contains the cls token present in ViT plus a special token, dist, used for distillation.

Parameters

emb_size (int) – Embedding dimensions

training: bool