{
  "_name_or_path": "/ceph_home/lizhuoqun2021/hf_models/llama-2-7b",
  "alibi": false,
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "fuse_attention_ffn": false,
  "fuse_attention_qkv": false,
  "fuse_sequence_parallel_allreduce": false,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 11008,
  "long_sequence_init_args": {},
  "long_sequence_strategy_name": null,
  "long_sequence_strategy_type": null,
  "max_position_embeddings": 4096,
  "mlp_bias": false,
  "model_type": "llama",
  "no_recompute_layers": null,
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 32,
  "pad_token_id": 0,
  "paddlenlp_version": null,
  "pp_recompute_interval": 1,
  "pretraining_tp": 1,
  "recompute_granularity": "full",
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_scaling_factor": 1.0,
  "rope_scaling_type": null,
  "rope_theta": 10000.0,
  "seq_length": 2048,
  "sequence_parallel": false,
  "tensor_parallel_output": true,
  "tie_word_embeddings": false,
  "transformers_version": "4.41.1",
  "use_fused_rms_norm": false,
  "use_fused_rope": false,
  "use_long_sequence_strategies": false,
  "use_recompute": false,
  "virtual_pp_degree": 1,
  "vocab_size": 32000
}