metadata
library_name: keras-hub
This is a SAM model uploaded using the KerasHub library and can be used with JAX, TensorFlow, and PyTorch backends.
This model is related to a ImageSegmenter task.
Model config:
- name: sam_backbone
- trainable: True
- image_encoder: {'module': 'keras_hub.src.models.vit_det.vit_det_backbone', 'class_name': 'ViTDetBackbone', 'config': {'name': 'vi_t_det_backbone', 'trainable': True, 'image_shape': [1024, 1024, 3], 'patch_size': 16, 'hidden_size': 1024, 'num_layers': 24, 'intermediate_dim': 4096, 'num_heads': 16, 'num_output_channels': 256, 'use_bias': True, 'use_abs_pos': True, 'use_rel_pos': True, 'window_size': 14, 'global_attention_layer_indices': [5, 11, 17, 23], 'layer_norm_epsilon': 1e-06}, 'registered_name': 'keras_hub>ViTDetBackbone'}
- prompt_encoder: {'module': 'keras_hub.src.models.sam.sam_prompt_encoder', 'class_name': 'SAMPromptEncoder', 'config': {'name': 'sam_prompt_encoder', 'trainable': True, 'dtype': {'module': 'keras', 'class_name': 'DTypePolicy', 'config': {'name': 'float32'}, 'registered_name': None}, 'hidden_size': 256, 'image_embedding_size': [64, 64], 'input_image_size': [1024, 1024], 'mask_in_channels': 16, 'activation': 'gelu'}, 'registered_name': 'keras_hub>SAMPromptEncoder'}
- mask_decoder: {'module': 'keras_hub.src.models.sam.sam_mask_decoder', 'class_name': 'SAMMaskDecoder', 'config': {'name': 'sam_mask_decoder', 'trainable': True, 'dtype': {'module': 'keras', 'class_name': 'DTypePolicy', 'config': {'name': 'float32'}, 'registered_name': None}, 'hidden_size': 256, 'num_layers': 2, 'intermediate_dim': 2048, 'num_heads': 8, 'embedding_dim': 256, 'num_multimask_outputs': 3, 'iou_head_depth': 3, 'iou_head_hidden_dim': 256, 'activation': 'gelu'}, 'registered_name': 'keras_hub>SAMMaskDecoder'}
This model card has been generated automatically and should be completed by the model author. See Model Cards documentation for more information.