config.json

{
    "alibi": false,
    "architectures": [
        "FalconForCausalLM"
    ],
    "attention_dropout": 0.0,
    "bias": false,
    "bos_token_id": 11,
    "eos_token_id": 11,
    "hidden_dropout": 0.0,
    "hidden_size": 14848,
    "initializer_range": 0.02,
    "layer_norm_epsilon": 1e-05,
    "model_type": "falcon",
    "multi_query": true,
    "new_decoder_architecture": true,
    "num_attention_heads": 232,
    "num_hidden_layers": 80,
    "num_kv_heads": 8,
    "parallel_attn": true,
    "torch_dtype": "bfloat16",
    "transformers_version": "4.32.0",
    "use_cache": true,
    "vocab_size": 65024,
    "pretraining_tp": 1,
    "pad_token_id": 0,
    "quantization_config": {
        "bits": 4,
        "group_size": 128,
        "damp_percent": 0.1,
        "desc_act": true,
        "sym": true,
        "true_sequential": true,
        "model_name_or_path": null,
        "model_file_base_name": "model",
        "quant_method": "gptq"
    }
}