{
  "architectures": [
    "DeepseekVLHybridModel"
  ],
  "dtype": "float32",
  "high_res_vision_config": {
    "attention_dropout": 0.0,
    "global_attn_indexes": [
      0
    ],
    "hidden_act": "gelu",
    "hidden_size": 16,
    "image_size": 128,
    "initializer_range": 1e-10,
    "intermediate_size": 37,
    "layer_norm_eps": 1e-06,
    "mlp_dim": 24,
    "mlp_ratio": 4.0,
    "model_type": "sam_vision_model",
    "num_attention_heads": 4,
    "num_channels": 3,
    "num_hidden_layers": 2,
    "num_pos_feats": 128,
    "output_channels": 4,
    "patch_size": 32,
    "qkv_bias": true,
    "scale": 8,
    "use_abs_pos": true,
    "use_rel_pos": true,
    "window_size": 14
  },
  "image_token_id": 98,
  "model_type": "deepseek_vl_hybrid",
  "text_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "bos_token_id": 100000,
    "eos_token_id": 100001,
    "head_dim": 4,
    "hidden_act": "silu",
    "hidden_size": 16,
    "initializer_range": 0.02,
    "intermediate_size": 37,
    "max_position_embeddings": 512,
    "mlp_bias": false,
    "model_type": "llama",
    "num_attention_heads": 4,
    "num_hidden_layers": 2,
    "num_key_value_heads": 4,
    "pad_token_id": 100001,
    "pretraining_tp": 1,
    "rms_norm_eps": 1e-06,
    "rope_parameters": {
      "rope_theta": 10000.0,
      "rope_type": "default"
    },
    "tie_word_embeddings": false,
    "use_cache": true,
    "vocab_size": 100016
  },
  "tie_word_embeddings": true,
  "transformers_version": "5.16.0.dev0",
  "vision_config": {
    "attention_dropout": 0.0,
    "hidden_act": "gelu",
    "hidden_size": 16,
    "image_size": 32,
    "intermediate_size": 37,
    "layer_norm_eps": 1e-06,
    "model_type": "siglip_vision_model",
    "num_attention_heads": 4,
    "num_channels": 3,
    "num_hidden_layers": 1,
    "patch_size": 8,
    "vision_use_head": false
  }
}