vision_tower/config.json

{
  "_name_or_path": "./checkpoints/describe_anything_v0.7.2_1ep_e2a_with_words_paug_samv2_lvisv4_mapillary_cocostuff_openimages_context_crop_constrained_blur0.5/vision_tower",
  "architectures": [
    "SiglipVisionModel"
  ],
  "attention_dropout": 0.0,
  "decouple_encoder_for_local_global": false,
  "dual_input_mode": null,
  "hidden_act": "gelu_pytorch_tanh",
  "hidden_size": 1152,
  "image_size": 384,
  "intermediate_size": 4304,
  "layer_norm_eps": 1e-06,
  "model_type": "siglip_vision_model",
  "num_attention_heads": 16,
  "num_channels": 3,
  "num_hidden_layers": 27,
  "num_mask_channels": 1,
  "patch_size": 14,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.36.2"
}