config.json

{
  "architectures": [
    "Model"
  ],
  "caption_model": "HuggingFaceTB/SmolLM-135M-Instruct",
  "image_model": "mobilenetv4_conv_medium.e500_r256_in1k",
  "model_type": "vlm",
  "num_projections": 3,
  "torch_dtype": "float32",
  "transformers_version": "4.42.3"
}