{
  "architectures": [
    "AXK2ForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "attention_output_gate": true,
  "attn_gate_fused": true,
  "aux_loss_alpha": 0.001,
  "bos_token_id": 163691,
  "dtype": "bfloat16",
  "eos_token_id": 163704,
  "ep_size": 2,
  "gated_norm": true,
  "gated_norm_rank": 16,
  "head_dim": 64,
  "hidden_act": "silu",
  "hidden_size": 7168,
  "index_head_dim": 128,
  "index_n_heads": 64,
  "index_topk": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 18432,
  "kv_lora_rank": 512,
  "layer_types": [
    "deepseek_sparse_attention",
    "deepseek_sparse_attention"
  ],
  "max_position_embeddings": 8192,
  "mlp_bias": false,
  "mlp_layer_types": [
    "dense",
    "sparse"
  ],
  "model_type": "axk2",
  "moe_intermediate_size": 2048,
  "n_group": 1,
  "n_routed_experts": 2,
  "n_shared_experts": 1,
  "norm_topk_prob": true,
  "num_attention_heads": 64,
  "num_experts_per_tok": 1,
  "num_hidden_layers": 2,
  "num_key_value_heads": 64,
  "num_nextn_predict_layers": 0,
  "pad_token_id": 163692,
  "pretraining_tp": 1,
  "q_lora_rank": 1536,
  "qk_head_dim": 192,
  "qk_nope_head_dim": 128,
  "qk_rope_head_dim": 64,
  "rms_norm_eps": 1e-06,
  "rope_parameters": {
    "beta_fast": 32.0,
    "beta_slow": 1.0,
    "factor": 2.0,
    "mscale": 1.0,
    "mscale_all_dim": 1.0,
    "original_max_position_embeddings": 4096,
    "rope_theta": 1000000,
    "rope_type": "yarn"
  },
  "routed_scaling_factor": 2.5,
  "scoring_func": "sigmoid",
  "seq_aux": true,
  "tie_word_embeddings": false,
  "topk_group": 1,
  "topk_method": "noaux_tc",
  "transformers_version": "5.15.0.dev0",
  "use_cache": true,
  "v_head_dim": 128,
  "vocab_size": 163840
}