{
  "_layer_implementation": "Qwen2MemDecoderLayer",
  "_ahn_implementation": "Mamba2",
  "_name_or_path": "Qwen/Qwen2.5-14B-Instruct",
  "architectures": [
    "Qwen2ForCausalLM"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "dy_num_attn_sinks": 128,
  "dy_sliding_window": 2048,
  "eos_token_id": 151645,
  "fa_layer_ids": "",
  "hidden_act": "silu",
  "hidden_size": 5120,
  "initializer_range": 0.02,
  "intermediate_size": 13824,
  "is_causal": true,
  "loss_type": [
    "kl"
  ],
  "max_position_embeddings": 32768,
  "max_window_layers": 70,
  "ahn_position": "random",
  "model_type": "qwen2",
  "num_attention_heads": 40,
  "num_hidden_layers": 48,
  "num_key_value_heads": 8,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000.0,
  "sliding_window": 256,
  "sliding_window_type": "random",
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.51.0",
  "use_cache": false,
  "use_layerwise_decay": false,
  "use_ahn_router": false,
  "use_normalized_l2": true,
  "use_q_proj": false,
  "use_sliding_window": false,
  "vocab_size": 152064
}