config.json

{
  "architectures": [
    "LightonOcrForConditionalGeneration"
  ],
  "dtype": "float32",
  "eos_token_id": 151645,
  "image_token_index": 151655,
  "model_type": "lighton_ocr",
  "multimodal_projector_bias": false,
  "pad_token_id": 151643,
  "projector_hidden_act": "gelu",
  "spatial_merge_size": 2,
  "text_config": {
    "architectures": [
      "Qwen3ForCausalLM"
    ],
    "attention_bias": false,
    "attention_dropout": 0,
    "head_dim": 8,
    "hidden_act": "silu",
    "hidden_size": 32,
    "initializer_range": 0.02,
    "intermediate_size": 64,
    "layer_types": [
      "full_attention",
      "full_attention"
    ],
    "max_position_embeddings": 512,
    "max_window_layers": 2,
    "model_type": "qwen3",
    "num_attention_heads": 4,
    "num_hidden_layers": 2,
    "num_key_value_heads": 2,
    "rms_norm_eps": 1e-06,
    "rope_scaling": null,
    "rope_theta": 10000,
    "sliding_window": null,
    "tie_word_embeddings": true,
    "use_cache": true,
    "use_qk_norm": true,
    "use_sliding_window": false,
    "vocab_size": 151936
  },
  "transformers_version": "4.57.6",
  "use_cache": false,
  "vision_config": {
    "attention_dropout": 0,
    "head_dim": 8,
    "hidden_act": "silu",
    "hidden_size": 32,
    "image_size": 56,
    "initializer_range": 0.02,
    "intermediate_size": 64,
    "model_type": "pixtral",
    "num_attention_heads": 4,
    "num_channels": 3,
    "num_hidden_layers": 2,
    "patch_size": 14,
    "rope_theta": 10000
  },
  "vision_feature_layer": -1
}