{
    "architectures": [
        "ApertusForCausalLM"
    ],
    "attention_bias": false,
    "attention_dropout": 0.0,
    "bos_token_id": 1,
    "dtype": "float32",
    "eos_token_id": [
        2,
        68,
        72
    ],
    "hidden_act": "xielu",
    "hidden_dropout": 0.0,
    "hidden_size": 4096,
    "initializer_range": 0.02,
    "intermediate_size": 21504,
    "max_position_embeddings": 65536,
    "mlp_bias": false,
    "model_type": "apertus",
    "num_attention_heads": 32,
    "num_hidden_layers": 32,
    "num_key_value_heads": 8,
    "pad_token_id": 3,
    "post_norm": false,
    "qk_norm": true,
    "quantization": {
        "group_size": 64,
        "bits": 6,
        "mode": "affine",
        "model.embed_tokens": {
            "bits": 6,
            "group_size": 64
        },
        "lm_head": {
            "bits": 6,
            "group_size": 64
        }
    },
    "quantization_config": {
        "group_size": 64,
        "bits": 6,
        "mode": "affine",
        "model.embed_tokens": {
            "bits": 6,
            "group_size": 64
        },
        "lm_head": {
            "bits": 6,
            "group_size": 64
        }
    },
    "rms_norm_eps": 1e-05,
    "rope_parameters": {
        "factor": 8.0,
        "high_freq_factor": 4.0,
        "low_freq_factor": 1.0,
        "original_max_position_embeddings": 8192,
        "rope_theta": 12000000,
        "rope_type": "llama3",
        "type": "llama3"
    },
    "rope_theta": 12000000,
    "tie_word_embeddings": false,
    "transformers_version": "5.8.1",
    "use_cache": false,
    "vocab_size": 131072
}