config.json

{
    "adapter_layer_norm_eps": 1e-06,
    "alignment_activation_fn": "swiglu",
    "alignment_intermediate_size": 28672,
    "architectures": [
        "AyaVisionForConditionalGeneration"
    ],
    "downsample_factor": 2,
    "image_token_index": 255036,
    "max_splits_per_img": 12,
    "model_type": "aya_vision",
    "projector_hidden_act": "gelu",
    "quantization": {
        "group_size": 64,
        "bits": 6
    },
    "text_config": {
        "hidden_size": 4096,
        "intermediate_size": 14336,
        "logit_scale": 0.25,
        "model_type": "cohere2",
        "num_attention_heads": 32,
        "num_hidden_layers": 32,
        "num_key_value_heads": 8,
        "rope_theta": 50000,
        "torch_dtype": "float16",
        "use_qk_norm": false
    },
    "torch_dtype": "float16",
    "transformers_version": "4.50.0.dev0",
    "vision_config": {
        "hidden_size": 1152,
        "image_size": 364,
        "intermediate_size": 4304,
        "model_type": "siglip_vision_model",
        "num_attention_heads": 16,
        "num_hidden_layers": 27,
        "patch_size": 14,
        "torch_dtype": "float16",
        "vision_use_head": false,
        "skip_vision": true
    },
    "vision_feature_layer": -1,
    "vision_feature_select_strategy": "full"
}