ijepa_base_16_cls(pretrained: bool = False, overrides: object = {})A linear probe on I-JEPA ViT-B/16.
Model Size
Parameters
pretrainedbool= FalseNo weights are published for this size;
True raises.**overridesobject= {}Optional
IJEPAConfig field overrides.Returns
IJEPAForImageClassificationThe pretraining networks plus a linear head over the frozen target encoder.
Notes
Reference: Assran et al., arXiv:2301.08243, Appendix A.2 — the paper's evaluation protocol is a linear probe on the average-pooled target encoder.
Examples
>>> from lucid.models import ijepa_base_16_cls
>>> model = ijepa_base_16_cls(
... image_size=32, patch_size=8, dim=16, depth=1, num_heads=2,
... predictor_dim=8, predictor_depth=1, num_classes=10, min_keep=1)
>>> model.config.num_classes
10