keras/dinov3_vit_small_plus_lvd1689m

This is a DINOV3 model uploaded using the KerasHub library and can be used with JAX, TensorFlow, and PyTorch backends. Model config:

• name: dinov3_backbone
• trainable: True
• dtype: {'module': 'keras', 'class_name': 'DTypePolicy', 'config': {'name': 'float32'}, 'registered_name': None}
• patch_size: 16
• num_layers: 12
• hidden_dim: 384
• num_heads: 6
• intermediate_dim: 1536
• num_register_tokens: 4
• use_mask_token: True
• layer_scale_init_value: 1.0
• hidden_activation: silu
• use_gated_mlp: True
• use_query_bias: True
• use_key_bias: False
• use_value_bias: True
• use_proj_bias: True
• use_mlp_bias: True
• attention_dropout: 0.0
• drop_path_rate: 0.0
• layer_norm_eps: 1e-05
• image_shape: [224, 224, 3]
• rope_theta: 100.0
• apply_layernorm: False

This model card has been generated automatically and should be completed by the model author. See Model Cards documentation for more information.