config.json

{
  "_name_or_path": "MLLMModel",
  "version": 2.6,
  "architectures": [
    "MLLMModel"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 3584,
  "initializer_range": 0.02,
  "intermediate_size": 18944,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "num_attention_heads": 28,
  "num_hidden_layers": 28,
  "num_key_value_heads": 4,
  "rms_norm_eps": 1e-06,
  "rope_theta": 1000000.0,
  "sliding_window": 131072,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.40.0",
  "use_cache": true,
  "use_sliding_window": false,
  "vocab_size": 151666,
  "batch_vision_input": true,
  "drop_vision_last_layer": false,
  "image_size": 448,
  "model_type": "mllm",
  "patch_size": 14,
  "query_num": 64,
  "slice_config": {
    "max_slice_nums": 9,
    "patch_size": 14,
    "model_type": "mllm"
  },
  "slice_mode": true,
  "vision_config": {
    "hidden_size": 1152,
    "image_size": 980,
    "intermediate_size": 4304,
    "model_type": "siglip",
    "num_attention_heads": 16,
    "num_hidden_layers": 27,
    "patch_size": 14
  }
}