{
  "architectures": [
    "DeepseekV4ForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 0,
  "eos_token_id": 1,
  "expert_dtype": "fp4",
  "hc_eps": 1e-06,
  "hc_mult": 4,
  "hc_sinkhorn_iters": 20,
  "head_dim": 512,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "index_head_dim": 128,
  "index_n_heads": 64,
  "index_topk": 512,
  "initializer_range": 0.02,
  "max_position_embeddings": 1048576,
  "model_type": "deepseek_v4",
  "moe_intermediate_size": 2048,
  "n_routed_experts": 256,
  "n_shared_experts": 1,
  "norm_topk_prob": true,
  "num_attention_heads": 64,
  "num_experts_per_tok": 6,
  "num_hidden_layers": 43,
  "num_hash_layers": 3,
  "num_key_value_heads": 1,
  "num_nextn_predict_layers": 1,
  "o_groups": 8,
  "o_lora_rank": 1024,
  "q_lora_rank": 1024,
  "qk_rope_head_dim": 64,
  "quantization_config": {
    "quant_method": "dsv4_mxfp4_int8",
    "format": "mxfp4_int8_packed",
    "config_groups": {
      "experts_w4a16": {
        "weights": {
          "num_bits": 4,
          "type": "float",
          "format": "mxfp4",
          "symmetric": false,
          "group_size": 32,
          "strategy": "group",
          "scale_mode": "native_e8m0",
          "scale_dtype": "e8m0"
        },
        "input_activations": {
          "num_bits": 16,
          "type": "float"
        },
        "targets": [
          "*.ffn.experts.routed_experts.w13_weight",
          "*.ffn.experts.routed_experts.w2_weight"
        ]
      },
      "linears_w8a16": {
        "weights": {
          "num_bits": 8,
          "type": "int",
          "symmetric": true,
          "strategy": "channel"
        },
        "input_activations": {
          "num_bits": 16,
          "type": "float"
        },
        "targets": [
          "*.attn.fused_wqa_wkv",
          "*.attn.wq_b",
          "*.attn.wo_a",
          "*.attn.wo_b",
          "*.attn.indexer.wq_b",
          "*.attn.indexer.compressor.wkv",
          "*.ffn.shared_experts.gate_up_proj",
          "*.ffn.shared_experts.down_proj"
        ]
      }
    },
    "ignore": [
      "embed",
      "head",
      "norm",
      "lm_head",
      "*norm.weight",
      "attn.attn_sink",
      "*.gate.*",
      "hc_*",
      "*.hc_attn_*",
      "*.hc_ffn_*"
    ]
  },
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "beta_fast": 32,
    "beta_slow": 1,
    "factor": 16,
    "original_max_position_embeddings": 65536,
    "type": "yarn"
  },
  "rope_theta": 10000,
  "routed_scaling_factor": 1.5,
  "scoring_func": "sqrtsoftplus",
  "sliding_window": 128,
  "swiglu_limit": 10.0,
  "tie_word_embeddings": false,
  "topk_method": "noaux_tc",
  "torch_dtype": "bfloat16",
  "transformers_version": "4.57.1",
  "use_cache": true,
  "vocab_size": 129280,
  "compress_rope_theta": 160000,
  "compress_ratios": [
    0,
    0,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    128,
    4,
    0
  ],
  "appmana": {
    "kernels": [
      "flash_mla.flash_sparse_mla_decode",
      "flash_mla.triton_sparse_int8_mla_decode",
      "vllm.models.deepseek_v4.nvidia_sm86.triton_kernels.sparse_attention_triton"
    ],
    "cache_type": "fp8_ds_mla"
  }
}