{
  "config_version": "1.0.0",
  "model_id": "meta-llama/Llama-3.1-70B",
  "model_kinds": [
    "ARTIFACT"
  ],
  "model_class": {
    "module": "furiosa_llm_models.llama3.symbolic.aramco_specdec",
    "name": "LlamaForCausalLM"
  },
  "llm_config": {
    "optimization_config": {
      "attention_type": "PAGED_ATTENTION",
      "optimize_rope": true,
      "optimize_packed": true,
      "decompose_layernorm": false,
      "optimize_furiosa": false,
      "use_unsplit_packed": false,
      "compact_causal_mask": false,
      "use_rngd_gelu": false,
      "causal_mask_free_decoding": true,
      "kv_cache_sharing_across_beams": false,
      "inbound_beamsearch_softmax": false,
      "calculate_logit_only_for_last_token": false,
      "optimized_for_speculative_decoding": true,
      "use_2d_masks": false,
      "merged_kv_indices": false
    },
    "quantization_config": {
      "weight": "int8",
      "activation": "bf16",
      "kv_cache": "bf16",
      "use_mcp": true
    }
  },
  "components_versions": {
    "furiosa_llm": {
      "version": "0.1.0-dev",
      "git_hash": "137c9e3",
      "build_time": null
    },
    "furiosa_ir": {
      "version": "0.11.0-dev",
      "git_hash": "4467f6a699",
      "build_time": "2025-08-28T02:41:28Z"
    },
    "furiosa_runtime": {
      "version": "2025.3.2",
      "git_hash": "137c9e383",
      "build_time": "2025-08-28T02:41:32Z"
    },
    "furiosa_model_compressor": {
      "version": "2025.3.0 (rev: 4cd9804)",
      "git_hash": null,
      "build_time": null
    }
  }
}