{
  "algorithm": "model_free_ptq RTN (data-free weight-only PTQ)",
  "audit": {
    "audited_before_atomic_publish": true,
    "checkpoint": ".",
    "elements_by_dtype": {
      "BF16": 8669145856,
      "F32": 35328,
      "F8_E4M3": 8682209280,
      "U8": 69457674240
    },
    "finite_positive_weight_global_scales": 26496,
    "format": "nvfp4-pack-quantized",
    "gate_up_global_scale_bitwise_matches": 8832,
    "index_total_size_matches": true,
    "input_global_scale_tensors": 0,
    "keys_by_dtype": {
      "BF16": 891,
      "F32": 26544,
      "F8_E4M3": 26496,
      "U8": 26496
    },
    "n_routed_experts": 184,
    "n_shared_experts": 1,
    "preserved_payload_sha256_matches": 939,
    "preserved_tensor_keys": 939,
    "quantized_routed_modules": 26496,
    "raw_tensor_bytes": 95478316544,
    "raw_tensor_gb": 95.478316544,
    "raw_tensor_gib": 88.92111158370972,
    "scheme": "NVFP4A16 (W4A16)",
    "tensor_keys": 80427,
    "weight_zero_point_tensors": 0
  },
  "bf16_parent": {
    "remote_commit": "dea73266ec31a2453d115f6efc622e35d216ef81",
    "repo_id": "Baekpica/Solar-Open2-148B-A15B-REAP-184E-BF16"
  },
  "completed_at_unix": 1784747809.6015775,
  "devices": [
    "cuda:0",
    "cuda:1",
    "cuda:2",
    "cuda:3",
    "cuda:4",
    "cuda:5"
  ],
  "elapsed_seconds": 218.6971311569214,
  "expected_quantized_modules": 26496,
  "expected_raw_output_bytes": 95478316544,
  "ignore": [
    "re:^model\\.layers\\.\\d+\\.self_attn\\..*$",
    "re:^model\\.layers\\.\\d+\\.mlp\\.shared_experts\\..*$",
    "re:^model\\.layers\\.\\d+\\.mlp\\.gate.*$",
    "model.embed_tokens",
    "lm_head"
  ],
  "inherited_bf16_publication_cleanup": {
    "after": [
      "chat_template.jinja",
      "config.json",
      "generation_config.json",
      "model.safetensors.index.json",
      "reap_metadata.json",
      "tokenizer.json",
      "tokenizer_config.json"
    ],
    "before": [
      "LICENSE",
      "README.md",
      "bf16_audit.json",
      "calibration_state_audit.json",
      "chat_template.jinja",
      "config.json",
      "generation_config.json",
      "hub_artifact_manifest.json",
      "model.safetensors.index.json",
      "reap_metadata.json",
      "tokenizer.json",
      "tokenizer_config.json"
    ],
    "removed": [
      "LICENSE",
      "README.md",
      "bf16_audit.json",
      "calibration_state_audit.json",
      "hub_artifact_manifest.json"
    ],
    "verified_bf16_manifest": true
  },
  "max_workers": 6,
  "ptq_staging_audit": {
    "checkpoint": "PTQ staging checkpoint (not published)",
    "elements_by_dtype": {
      "BF16": 8669145856,
      "F32": 35328,
      "F8_E4M3": 8682209280,
      "U8": 69457674240
    },
    "finite_positive_weight_global_scales": 26496,
    "format": "nvfp4-pack-quantized",
    "gate_up_global_scale_bitwise_matches": 8832,
    "index_total_size_matches": true,
    "input_global_scale_tensors": 0,
    "keys_by_dtype": {
      "BF16": 891,
      "F32": 26544,
      "F8_E4M3": 26496,
      "U8": 26496
    },
    "n_routed_experts": 184,
    "n_shared_experts": 1,
    "preserved_payload_sha256_matches": 939,
    "preserved_tensor_keys": 939,
    "quantized_routed_modules": 26496,
    "raw_tensor_bytes": 95478316544,
    "raw_tensor_gb": 95.478316544,
    "raw_tensor_gib": 88.92111158370972,
    "scheme": "NVFP4A16 (W4A16)",
    "tensor_keys": 80427,
    "weight_zero_point_tensors": 0
  },
  "repack": {
    "all_tensor_payload_sha256_matches": 80427,
    "audited_before_atomic_publish": true,
    "header_and_prefix_bytes": 10504816,
    "input_global_scale_tensors": 0,
    "logical_tensor_payload_sha256": "27bf03cfe421a8137f011a3868af1ea95b12c38a2073b2df07058635cba67f0c",
    "max_physical_shard_bytes": 9999808584,
    "model_audit": {
      "audited_before_atomic_publish": true,
      "checkpoint": ".",
      "elements_by_dtype": {
        "BF16": 8669145856,
        "F32": 35328,
        "F8_E4M3": 8682209280,
        "U8": 69457674240
      },
      "finite_positive_weight_global_scales": 26496,
      "format": "nvfp4-pack-quantized",
      "gate_up_global_scale_bitwise_matches": 8832,
      "index_total_size_matches": true,
      "input_global_scale_tensors": 0,
      "keys_by_dtype": {
        "BF16": 891,
        "F32": 26544,
        "F8_E4M3": 26496,
        "U8": 26496
      },
      "n_routed_experts": 184,
      "n_shared_experts": 1,
      "preserved_payload_sha256_matches": 939,
      "preserved_tensor_keys": 939,
      "quantized_routed_modules": 26496,
      "raw_tensor_bytes": 95478316544,
      "raw_tensor_gb": 95.478316544,
      "raw_tensor_gib": 88.92111158370972,
      "scheme": "NVFP4A16 (W4A16)",
      "tensor_keys": 80427,
      "weight_zero_point_tensors": 0
    },
    "nonweight_files_preserved": 6,
    "output": ".",
    "physical_shard_bytes": [
      9999721088,
      9997516324,
      9964081700,
      9998226852,
      9999808080,
      9997515972,
      9999808584,
      9954241344,
      9999544392,
      5578357024
    ],
    "physical_size_cap": 10000000000,
    "raw_tensor_bytes": 95478316544,
    "raw_tensor_gib": 88.92111158370972,
    "rust_safe_open_shards": 10,
    "shard_count": 10,
    "source": "PTQ staging checkpoint (not published)",
    "tensor_keys": 80427,
    "total_physical_shard_bytes": 95488821360
  },
  "schema_version": 1,
  "scheme": "NVFP4A16",
  "solar_gate_up_fused_mapping": {
    "name_set": [
      "re:^model\\.layers\\.\\d+\\.mlp\\.experts\\.\\d+\\.gate_proj\\.weight$",
      "re:^model\\.layers\\.\\d+\\.mlp\\.experts\\.\\d+\\.up_proj\\.weight$"
    ],
    "partners": [
      "{prefix}.up_proj.weight"
    ],
    "primary_pattern": "^(?P<prefix>model\\.layers\\.\\d+\\.mlp\\.experts\\.\\d+)\\.gate_proj\\.weight$",
    "self_test_pairs": 1,
    "validation_adapter": "solar_source_shard_validation_v1"
  },
  "solar_gate_up_source_preflight": {
    "cross_shard_pairs_relocated": 4416,
    "gate_up_pairs": 8832,
    "relocated_gate_up_pairs": 8832,
    "relocated_gate_up_pairs_matched": 8832,
    "relocated_pair_names_exactly_once": 17664,
    "source_shards": 73,
    "source_shards_with_jobs": 73
  },
  "versions": {
    "PyYAML": "6.0.3",
    "compressed-tensors": "0.17.1",
    "huggingface-hub": "1.24.0",
    "llmcompressor": "0.12.0",
    "torch": "2.12.0+cu130",
    "transformers": "5.10.1"
  }
}