{ | |
"architectures": [ | |
"CoMPSiglipVisionModel" | |
], | |
"deepstack_type": "v1", | |
"global_interval": null, | |
"initializer_factor": 1.0, | |
"llm_hidden_size": 896, | |
"model_type": "siglip", | |
"spatial_merge_size": 2, | |
"temporal_patch_size": 1, | |
"torch_dtype": "float32", | |
"transformers_version": "4.47.0.dev0", | |
"use_deepstack": false, | |
"use_rope": true, | |
"vision_config": { | |
"hidden_size": 1152, | |
"image_size": 384, | |
"intermediate_size": 4304, | |
"model_type": "siglip_vision_model", | |
"num_attention_heads": 16, | |
"num_hidden_layers": 27, | |
"patch_size": 14, | |
"spatial_merge_size": 2, | |
"use_rope": true | |
} | |
} | |