{
  "architectures": [
    "Qwen2ForCausalLM"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "eos_token_id": 151643,
  "hidden_act": "silu",
  "hidden_size": 896,
  "initializer_range": 0.02,
  "intermediate_size": 4864,
  "max_position_embeddings": 131072,
  "max_window_layers": 24,
  "mha2mla": {
    "is_baseline": false,
    "is_gqa2mha2mla": false,
    "is_mla_from_scratch": false,
    "low_rank": 16,
    "model_name_or_path": "/cpfs01/shared/llm_ddd/jitao/code/MLA-FT/checkpoints/Qwen/Qwen2-0.5B",
    "partial_rope_version": "2-norm",
    "qk_tensor_path": "utils/qwen2_0.5B-2_norm_rank.pth",
    "rope_dim_for_mla": 8,
    "svd_init_method": "joint",
    "uniform_start_point": 0
  },
  "model_type": "qwen2",
  "num_attention_heads": 14,
  "num_hidden_layers": 24,
  "num_key_value_heads": 2,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000.0,
  "sliding_window": 131072,
  "tie_word_embeddings": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.51.0",
  "use_cache": true,
  "use_sliding_window": false,
  "vocab_size": 151936
}