config.json

{
  "_attn_implementation": "flash_attention_2",
  "_commit_hash": "3ab0a536b1778c885624ae920cd73d06228ae441",
  "architectures": [
    "Eagle2_5_VLForConditionalGeneration"
  ],
  "auto_map": {
    "AutoConfig": "nvidia/Eagle2-2B--configuration_eagle2_5_vl.Eagle2_5_VLConfig",
    "AutoModel": "nvidia/Eagle2-2B--modeling_eagle2_5_vl.Eagle2_5_VLForConditionalGeneration"
  },
  "downsample_ratio": 0.5,
  "dynamic_image_size": true,
  "force_image_size": 448,
  "image_token_index": 151667,
  "initializer_range": 0.02,
  "loss_version": "v1",
  "max_dynamic_tiles": 1,
  "min_dynamic_tiles": 1,
  "mlp_checkpoint": false,
  "model_type": "eagle_2_5_vl",
  "pad2square": false,
  "pre_feature_reduction": false,
  "select_layer": -1,
  "template": "qwen2-chat",
  "text_config": {
    "_attn_implementation_autoset": true,
    "_name_or_path": "./pretrained/Qwen2_5-1_5B-Instruct",
    "architectures": [
      "Qwen2ForCausalLM"
    ],
    "attention_dropout": 0.0,
    "bos_token_id": 151643,
    "eos_token_id": 151645,
    "hidden_act": "silu",
    "hidden_size": 1536,
    "initializer_range": 0.02,
    "intermediate_size": 8960,
    "max_position_embeddings": 32768,
    "max_window_layers": 21,
    "model_type": "qwen2",
    "num_attention_heads": 12,
    "num_hidden_layers": 28,
    "num_key_value_heads": 2,
    "rms_norm_eps": 1e-06,
    "rope_scaling": null,
    "rope_theta": 1000000.0,
    "sliding_window": 32768,
    "tie_word_embeddings": true,
    "torch_dtype": "bfloat16",
    "use_cache": false,
    "use_sliding_window": false,
    "vocab_size": 151674
  },
  "torch_dtype": "bfloat16",
  "transformers_version": "4.51.0",
  "use_backbone_lora": 0,
  "use_llm_lora": 0,
  "use_thumbnail": true,
  "vision_config": {
    "_attn_implementation_autoset": true,
    "attention_dropout": 0.0,
    "drop_path_rate": 0.1,
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 1152,
    "image_size": 448,
    "intermediate_size": 4304,
    "layer_norm_eps": 1e-06,
    "model_type": "siglip_vision_model",
    "num_attention_heads": 16,
    "num_channels": 3,
    "num_hidden_layers": 27,
    "patch_size": 14,
    "torch_dtype": "bfloat16"
  }
}