config.json

{
    "architectures": [
        "GemmaForCausalLM"
    ],
    "attention_bias": false,
    "attention_dropout": 0.0,
    "bos_token_id": 2,
    "eos_token_id": 1,
    "head_dim": 256,
    "hidden_activation": "gelu_pytorch_tanh",
    "hidden_size": 3072,
    "initializer_range": 0.02,
    "intermediate_size": 24576,
    "max_position_embeddings": 8192,
    "model_type": "gemma",
    "num_attention_heads": 16,
    "num_hidden_layers": 28,
    "num_key_value_heads": 16,
    "pad_token_id": 0,
    "rms_norm_eps": 1e-06,
    "rope_theta": 10000.0,
    "torch_dtype": "bfloat16",
    "transformers_version": "4.38.1",
    "use_cache": true,
    "vocab_size": 256000,
    "quantization_config": {
        "quant_method": "exl2",
        "version": "0.0.17",
        "bits": 6.0,
        "head_bits": 6,
        "calibration": {
            "rows": 100,
            "length": 2048,
            "dataset": "(default)"
        }
    }
}