config.json

{
  "_name_or_path": "nyunai/nyun-c2-llama3-61B",
  "architectures": [
    "LlamaForCausalLM"
  ],
  "auto_map": {
    "AutoConfig": "configuration_llama.LlamaConfig",
    "AutoModelForCausalLM": "modeling_llama.LlamaForCausalLM"
  },
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "hidden_act": "silu",
  "hidden_size": 8192,
  "initializer_range": 0.02,
  "intermediate_size": [
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    28672,
    13120,
    6421,
    13147,
    5797,
    6510,
    5491,
    8320,
    7594,
    9199,
    4712,
    4590,
    8467,
    5990,
    28343,
    5626,
    9352,
    8924,
    28672,
    28672,
    28672
  ],
  "max_position_embeddings": 8192,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": [
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    64,
    8,
    24,
    40,
    64,
    16,
    16,
    16,
    32,
    24,
    48,
    64,
    64,
    64,
    24,
    64,
    64,
    40,
    32,
    32,
    24
  ],
  "num_hidden_layers": 80,
  "num_key_value_heads": [
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    8,
    1,
    3,
    5,
    8,
    2,
    2,
    2,
    4,
    3,
    6,
    8,
    8,
    8,
    3,
    8,
    8,
    5,
    4,
    4,
    3
  ],
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "float16",
  "transformers_version": "4.41.2",
  "use_cache": true,
  "vocab_size": 128256,
  "first_compressed_layer_idx": 60
}