vjepa2_vit_large_cls(pretrained: bool = False, overrides: object = {})V-JEPA 2 ViT-L/16 under the paper's attentive probe.
Model Size
Parameters
pretrainedbool= FalseThe probe's own weights are not published;
True raises. The
backbone tags are reachable through vjepa2_vit_large.**overridesobject= {}Optional
VJEPA2Config field overrides.Returns
VJEPA2ForVideoClassificationThe pretraining networks, the attentive pooler and a classifier.
Notes
Reference: Assran et al., arXiv:2506.09985, 2025. The released
classifiers read a frozen backbone through three self-attention
blocks and one learned query, which is what num_pooler_layers
defaults to; num_classes defaults to 400, Kinetics-400's count.
Examples
>>> from lucid.models import AutoConfig
>>> AutoConfig.from_pretrained("vjepa2_vit_large_cls").num_classes
400