{
  "architectures": [
    "GptOssForCausalLM"
  ],
  "attention_bias": true,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoModel": "modeling_gpt_oss.GptOssModel",
    "AutoModelForCausalLM": "modeling_gpt_oss.GptOssForCausalLM"
  },
  "eos_token_id": 200002,
  "experts_per_token": 4,
  "head_dim": 64,
  "hidden_act": "silu",
  "hidden_size": 2880,
  "initial_context_length": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 2880,
  "layer_types": [
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "full_attention"
  ],
  "max_position_embeddings": 131072,
  "model_type": "gpt_oss",
  "num_attention_heads": 64,
  "num_experts_per_tok": 4,
  "num_hidden_layers": 24,
  "num_key_value_heads": 8,
  "num_local_experts": 32,
  "output_router_logits": false,
  "pad_token_id": 199999,
  "quantization_config": {
    "autoround_version": "0.6.1.dev",
    "backend": "auto",
    "bits": 4,
    "data_type": "int",
    "extra_config": {
      "model.layers.0.mlp.router.router": {
        "bits": 16
      },
      "model.layers.0.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.0.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.0.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.0.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.1.mlp.router.router": {
        "bits": 16
      },
      "model.layers.1.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.1.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.1.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.1.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.10.mlp.router.router": {
        "bits": 16
      },
      "model.layers.10.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.10.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.10.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.10.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.11.mlp.router.router": {
        "bits": 16
      },
      "model.layers.11.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.11.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.11.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.11.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.12.mlp.router.router": {
        "bits": 16
      },
      "model.layers.12.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.12.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.12.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.12.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.13.mlp.router.router": {
        "bits": 16
      },
      "model.layers.13.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.13.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.13.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.13.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.14.mlp.router.router": {
        "bits": 16
      },
      "model.layers.14.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.14.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.14.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.14.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.15.mlp.router.router": {
        "bits": 16
      },
      "model.layers.15.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.15.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.15.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.15.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.16.mlp.router.router": {
        "bits": 16
      },
      "model.layers.16.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.16.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.16.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.16.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.17.mlp.router.router": {
        "bits": 16
      },
      "model.layers.17.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.17.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.17.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.17.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.18.mlp.router.router": {
        "bits": 16
      },
      "model.layers.18.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.18.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.18.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.18.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.19.mlp.router.router": {
        "bits": 16
      },
      "model.layers.19.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.19.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.19.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.19.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.2.mlp.router.router": {
        "bits": 16
      },
      "model.layers.2.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.2.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.2.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.2.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.20.mlp.router.router": {
        "bits": 16
      },
      "model.layers.20.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.20.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.20.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.20.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.21.mlp.router.router": {
        "bits": 16
      },
      "model.layers.21.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.21.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.21.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.21.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.22.mlp.router.router": {
        "bits": 16
      },
      "model.layers.22.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.22.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.22.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.22.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.23.mlp.router.router": {
        "bits": 16
      },
      "model.layers.23.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.23.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.23.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.23.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.3.mlp.router.router": {
        "bits": 16
      },
      "model.layers.3.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.3.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.3.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.3.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.4.mlp.router.router": {
        "bits": 16
      },
      "model.layers.4.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.4.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.4.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.4.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.5.mlp.router.router": {
        "bits": 16
      },
      "model.layers.5.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.5.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.5.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.5.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.6.mlp.router.router": {
        "bits": 16
      },
      "model.layers.6.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.6.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.6.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.6.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.7.mlp.router.router": {
        "bits": 16
      },
      "model.layers.7.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.7.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.7.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.7.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.8.mlp.router.router": {
        "bits": 16
      },
      "model.layers.8.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.8.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.8.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.8.self_attn.v_proj": {
        "bits": 16
      },
      "model.layers.9.mlp.router.router": {
        "bits": 16
      },
      "model.layers.9.self_attn.k_proj": {
        "bits": 16
      },
      "model.layers.9.self_attn.o_proj": {
        "bits": 16
      },
      "model.layers.9.self_attn.q_proj": {
        "bits": 16
      },
      "model.layers.9.self_attn.v_proj": {
        "bits": 16
      }
    },
    "group_size": 128,
    "iters": 1000,
    "nsamples": 512,
    "packing_format": "auto_round:auto_gptq",
    "quant_method": "auto-round",
    "sym": true
  },
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "beta_fast": 32.0,
    "beta_slow": 1.0,
    "factor": 32.0,
    "original_max_position_embeddings": 4096,
    "rope_type": "yarn",
    "truncate": false
  },
  "rope_theta": 150000,
  "router_aux_loss_coef": 0.9,
  "sliding_window": 128,
  "swiglu_limit": 7.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.55.0",
  "use_cache": true,
  "vocab_size": 201088
}