config.json

{
    "architectures": [
      "LlamaForCausalLM"
    ],
    "bos_token_id": 1,
    "eos_token_id": 2,
    "hidden_act": "silu",
    "hidden_size": 2048,
    "initializer_range": 0.02,
    "intermediate_size": 5760,
    "max_position_embeddings": 4096,
    "model_type": "llama",
    "num_attention_heads": 16,
    "num_hidden_layers": 1,
    "num_key_value_heads": 4,
    "pretraining_tp": 1,
    "rms_norm_eps": 1e-05,
    "rope_scaling": null,
    "tie_word_embeddings": true,
    "torch_dtype": "bfloat16",
    "transformers_version": "4.41",
    "use_cache": true,
    "vocab_size": 73440,
    "head_dim": 128,
    "scale_emb": 12,
    "scale_width": 8,
    "rm_first_layer_preln": true
}