DINOv2

DINOv2#

class stable_pretraining.methods.DINOv2(encoder_name: str | Module = 'vit_small_patch16_224', projector_hidden_dim: int = 2048, projector_bottleneck_dim: int = 256, n_cls_prototypes: int = 65536, n_patch_prototypes: int = 8192, mask_ratio: float = 0.3, patch_loss_weight: float = 1.0, temperature_student: float = 0.1, temperature_teacher: float = 0.07, ema_decay_start: float = 0.996, ema_decay_end: float = 1.0, image_size: int = 224, pretrained: bool = False)[source]#

Bases: Module

DINOv2: DINO + iBOT with Sinkhorn-Knopp on CLS and patch prototypes.

Parameters:

  • encoder_name – timm ViT name (default "vit_small_patch16_224").

  • projector_hidden_dim – Hidden dim for both heads (default 2048).

  • projector_bottleneck_dim – Bottleneck dim (default 256).

  • n_cls_prototypes – CLS prototypes (default 65536).

  • n_patch_prototypes – Patch prototypes (default 8192).

  • mask_ratio – Patch mask ratio for the student (default 0.3).

  • patch_loss_weight – Weight on the patch loss (default 1.0).

  • temperature_student – Student softmax temperature (default 0.1).

  • temperature_teacher – Teacher temperature (default 0.07).

  • ema_decay_start – Initial backbone/head EMA (default 0.996).

  • ema_decay_end – Final EMA (default 1.0).

  • image_size – Input size (default 224).

  • pretrained – Load pretrained timm weights.

forward(global_views: Sequence[Tensor] | None = None, local_views: Sequence[Tensor] | None = None, images: Tensor | None = None) → DINOv2Output[source]#

Same as torch.nn.Module.forward().

Parameters:

  • *args – Whatever you decide to pass into the forward method.

  • **kwargs – Keyword arguments are also possible.

Returns:

Your model’s output