config.json

{
  "_name_or_path": "/share/xingxingrun/prune350m_to_yulong/output_350_125_32gpu_ft/checkpoint-500000",
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoModel": "inverse_llama2.LlamaModel",
    "AutoModelForCausalLM": "inverse_llama2.LlamaForCausalLM"
  },
  "bos_token_id": 0,
  "eos_token_id": 0,
  "head_dim": 64,
  "hidden_act": "silu",
  "hidden_size": 757,
  "hidden_size_in": 320,
  "initializer_range": 0.02,
  "intermediate_size": 966,
  "kv_dims": [
    192,
    256,
    256,
    256,
    256,
    256,
    256,
    256,
    192,
    256,
    192,
    256,
    192,
    192,
    256,
    192,
    256,
    192,
    192,
    256,
    192,
    192,
    256,
    192,
    192,
    192,
    192,
    192,
    192,
    320,
    256,
    192
  ],
  "m_start": 250,
  "max_position_embeddings": 2048,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 5,
  "num_hidden_layers": 32,
  "num_key_value_heads": 5,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "step_quota": 9,
  "tie_word_embeddings": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.42.0",
  "use_cache": true,
  "vocab_size": 49152
}