Walkowicz/kairos-sotos-1.0-14B

hf Qwen/Qwen3-14B 1 source, 11 claims · Watch

What each source says

Property Source Said Means here

Author

author

GGUF quantisations

1bit-MONSTER

receipt

Source: GGUF quantisations
Its words: 1bit-MONSTER
Read by: field:author
Said since: 2026-09-28 11:46 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6ab7c98640a9cf563795b362",
  "author": "1bit-MONSTER",
  "cardData": {
    "base_model": [
      "Qwen/Qwen3-14B"
    ],
    "license": "apache-2.0",
    "tags": [
      "gguf"
    ]
  },
  "createdAt": "2026-09-26T13:32:54.000Z",
  "downloads": 52,
  "gated": false,
  "id": "1bit-MONSTER/Qwen3-14B-GGUF",
  "lastModified": "2026-09-26T13:33:22.000Z",
  "likes": 0,
  "modelId": "1bit-MONSTER/Qwen3-14B-GGUF",
  "private": false,
  "sha": "35bd73c5af9de2073e03339e86c059c236519abb",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Qwen3-14B-Q4_K_M.gguf"
    },
    {
      "rfilename": "README.md"
    }
  ],
  "tags": [
    "gguf",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Author

author

Irfanuruchi

receipt

Source: GGUF quantisations
Its words: Irfanuruchi
Read by: field:author
Said since: 2026-10-04 12:16 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6ac22ed17eb43e9e3e0fa3f4",
  "author": "Irfanuruchi",
  "cardData": {
    "base_model": [
      "Qwen/Qwen3-14B"
    ],
    "datasets": [
      "Irfanuruchi/buildeng-v8-3b"
    ],
    "language": [
      "en"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "qwen3",
      "qwen",
      "gguf",
      "llama-cpp",
      "building-engineering",
      "hvac",
      "engineering",
      "fine-tuned"
    ]
  },
  "createdAt": "2026-10-04T10:47:45.000Z",
  "downloads": 126,
  "gated": false,
  "id": "Irfanuruchi/Qwen3-14B-BuildEng-GGUF",
  "lastModified": "2026-10-04T12:55:26.000Z",
  "likes": 0,
  "modelId": "Irfanuruchi/Qwen3-14B-BuildEng-GGUF",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "6b29c423269d065b3974c5b91ce60c7065d3cafc",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "PUBLICATION-SHA256SUMS.txt"
    },
    {
      "rfilename": "Qwen3-14B-BuildEng-Q4_K_M.gguf"
    },
    {
      "rfilename": "Qwen3-14B-BuildEng-Q5_K_M.gguf"
    },
    {
      "rfilename": "Qwen3-14B-BuildEng-Q8_0.gguf"
    },
    {
      "rfilename": "README.md"
    }
  ],
  "tags": [
    "gguf",
    "qwen3",
    "qwen",
    "llama-cpp",
    "building-engineering",
    "hvac",
    "engineering",
    "fine-tuned",
    "text-generation",
    "en",
    "dataset:Irfanuruchi/buildeng-v8-3b",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Author

author

KitsuneOff

receipt

Source: GGUF quantisations
Its words: KitsuneOff
Read by: field:author
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6abacb4e87cf5463185425f5",
  "author": "KitsuneOff",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "license": "apache-2.0",
    "license_link": "https://huggingface.co/Qwen/Qwen3-14B-GGUF/blob/main/LICENSE",
    "pipeline_tag": "text-generation"
  },
  "createdAt": "2026-09-28T20:17:18.000Z",
  "downloads": 115,
  "gated": false,
  "id": "KitsuneOff/Qwen3-14B-GGUF",
  "lastModified": "2026-09-28T20:17:21.000Z",
  "likes": 0,
  "modelId": "KitsuneOff/Qwen3-14B-GGUF",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "218e5a18729725edbfd0e3e2355d98ad8684a7f4",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "Qwen3-14B-Q4_K_M.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Q5_0.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Q5_K_M.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Q6_K.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Q8_0.gguf"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "params"
    }
  ],
  "tags": [
    "gguf",
    "text-generation",
    "arxiv:2309.00071",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Author

author

Om22s

receipt

Source: GGUF quantisations
Its words: Om22s
Read by: field:author
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aabb0e05ed906a7bbc3ec17",
  "author": "Om22s",
  "cardData": {
    "base_model": [
      "Qwen/Qwen3-14B",
      "Qwen/Qwen3.6-35B-A3B",
      "Qwen/Qwen3.8-27B",
      "Qwen/Qwen3-8B"
    ],
    "base_model_relation": "adapter",
    "language": [
      "en"
    ],
    "library_name": "peft",
    "license": "apache-2.0",
    "model-index": [
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3-14b-rightsclean-attrv1)",
        "results": [
          {
            "dataset": {
              "name": "Alexandria four-book product fixture (768 paired rows, temperature 0, batch 25, default prompt, reasoning off)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 61.7)",
                "type": "accuracy",
                "value": 73.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "Alexandria four-book product fixture, reasoning low budget 1024, JSON schema",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 66.1)",
                "type": "accuracy",
                "value": 74.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), reasoning low budget 1024",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 68.9)",
                "type": "accuracy",
                "value": 75.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.9)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine held-out PDNC novels, 2,655 rows, IQ3_XXS, reasoning low, michel2_full (raw; 90.9 with the roster-line echo read as the app reads it)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.6)",
                "type": "accuracy",
                "value": 71.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q3_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.4)",
                "type": "accuracy",
                "value": 99.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.2)",
                "type": "accuracy",
                "value": 92.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.9)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine held-out PDNC novels, 2,655 rows, IQ3_XXS, reasoning low, michel2_full (raw; 90.9 with the roster-line echo read as the app reads it)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.6)",
                "type": "accuracy",
                "value": 71.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3-8b-rightsclean-window25-michel2v2)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning low, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 77.0)",
                "type": "accuracy",
                "value": 82.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-8B Q4_K_M, RX 9070 XT",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), Q4_K_M, reasoning low, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 76.6)",
                "type": "accuracy",
                "value": 82.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-8B Q4_K_M, RX 9070 XT",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2)",
        "results": [
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ1_M, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.7)",
                "type": "accuracy",
                "value": 92.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ2_XXS, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.9)",
                "type": "accuracy",
                "value": 91.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ3_XXS, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 92.5)",
                "type": "accuracy",
                "value": 92.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-Q4_K_XL, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 93.2)",
                "type": "accuracy",
                "value": 91.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      }
    ],
    "pipeline_tag": "text-generation",
    "tags": [
      "lora",
      "peft",
      "speaker-attribution",
      "audiobook",
      "qwen3",
      "rights-reviewed"
    ]
  },
  "createdAt": "2026-09-17T09:20:32.000Z",
  "downloads": 165,
  "gated": false,
  "id": "Om22s/alexandria-qwen3-attribution",
  "lastModified": "2026-09-29T07:14:41.000Z",
  "library_name": "peft",
  "likes": 0,
  "modelId": "Om22s/alexandria-qwen3-attribution",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "f54a4cd9599f4f496eae474cf9f6daee40b6af22",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/qwen3-14b-rightsclean-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/qwen3-14b-rightsclean-seed2-attrv2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/qwen3-14b-riqua3ep-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/qwen3-14b-riquax2-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/README.md"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/qwen3-8b-rightsclean-window25-michel2v2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/README.md"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/qwen3.6-35b-a3b-rightsclean-michel2v1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/README.md"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/qwen3.6-35b-a3b-rightsclean-window25-michel2v2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/README.md"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/qwen3.8-27b-rightsclean-michel2v1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/training_manifest.json"
    },
    {
      "rfilename": "default_prompts_attribute.txt"
    },
    {
      "rfilename": "examples/infer.py"
    },
    {
      "rfilename": "examples/serve_llamacpp.sh"
    },
    {
      "rfilename": "prompts/attrv1.txt"
    },
    {
      "rfilename": "prompts/attrv2.txt"
    },
    {
      "rfilename": "prompts/michel2v1.txt"
    },
    {
      "rfilename": "prompts/michel2v2.txt"
    }
  ],
  "tags": [
    "peft",
    "safetensors",
    "gguf",
    "lora",
    "speaker-attribution",
    "audiobook",
    "qwen3",
    "rights-reviewed",
    "text-generation",
    "en",
    "base_model:Qwen/Qwen3-14B",
    "base_model:adapter:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "model-index",
    "region:us"
  ]
}

—

Author

author

Walkowicz

receipt

Source: GGUF quantisations
Its words: Walkowicz
Read by: field:author
Said since: 2026-09-28 11:46 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6ab2aa765a05167560ef154e",
  "author": "Walkowicz",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en",
      "pt"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "gguf",
      "qwen3",
      "text-generation"
    ]
  },
  "createdAt": "2026-09-22T16:19:02.000Z",
  "downloads": 25,
  "gated": false,
  "id": "Walkowicz/kairos-sotos-1.0-14B",
  "lastModified": "2026-09-22T19:47:32.000Z",
  "likes": 0,
  "modelId": "Walkowicz/kairos-sotos-1.0-14B",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "d1c97e5d75be337e86ad69ce35cccb9c4660625a",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Modelfile"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "kairos-logo.png"
    },
    {
      "rfilename": "kairos-sotos-1.0-14B-Q6_K.gguf"
    }
  ],
  "tags": [
    "gguf",
    "qwen3",
    "text-generation",
    "en",
    "pt",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Author

author

anyalkonsafe

receipt

Source: GGUF quantisations
Its words: anyalkonsafe
Read by: field:author
Said since: 2026-10-05 06:21 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6ac3105fea36a8977d078842",
  "author": "anyalkonsafe",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en"
    ],
    "library_name": "peft",
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "qwen",
      "qwen3",
      "gguf",
      "lora",
      "qlora",
      "unsloth",
      "persona",
      "tool-calling",
      "text-generation"
    ]
  },
  "createdAt": "2026-10-05T02:50:07.000Z",
  "downloads": 20,
  "gated": false,
  "id": "anyalkonsafe/evilguy-3-gguf",
  "lastModified": "2026-10-05T02:55:33.000Z",
  "library_name": "peft",
  "likes": 0,
  "modelId": "anyalkonsafe/evilguy-3-gguf",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "f130250fe6b4141bd51dc886512ce57157068643",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "evilguy-3-q4_k_m.gguf"
    }
  ],
  "tags": [
    "peft",
    "gguf",
    "qwen",
    "qwen3",
    "lora",
    "qlora",
    "unsloth",
    "persona",
    "tool-calling",
    "text-generation",
    "en",
    "base_model:Qwen/Qwen3-14B",
    "base_model:adapter:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Author

author

chenhaodev

receipt

Source: GGUF quantisations
Its words: chenhaodev
Read by: field:author
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6a9b8f6727087a68413ae2cc",
  "author": "chenhaodev",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "zh"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "medical",
      "internal-medicine",
      "qwen",
      "gguf",
      "llama.cpp",
      "chinese"
    ]
  },
  "createdAt": "2026-09-05T03:41:27.000Z",
  "downloads": 248,
  "gated": false,
  "id": "chenhaodev/qwen3-14b-internal-med-gguf",
  "lastModified": "2026-09-05T05:00:01.000Z",
  "likes": 0,
  "modelId": "chenhaodev/qwen3-14b-internal-med-gguf",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "cc9ac7828cc7c901b54673a4d73e9b7d64425779",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Qwen3-14B-Internal-Med-q4_k_m.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Internal-Med-q5_k_m.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Internal-Med-q8_0.gguf"
    },
    {
      "rfilename": "README.md"
    }
  ],
  "tags": [
    "gguf",
    "medical",
    "internal-medicine",
    "qwen",
    "llama.cpp",
    "chinese",
    "text-generation",
    "zh",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Author

author

selorahomes

receipt

Source: GGUF quantisations
Its words: selorahomes
Read by: field:author
Said since: 2026-10-02 12:01 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6abbf79c51f0022e9a4225f9",
  "author": "selorahomes",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "text-generation",
      "qwen",
      "qwen3",
      "gguf",
      "home-assistant",
      "home-automation",
      "smart-home",
      "ollama"
    ]
  },
  "createdAt": "2026-09-29T17:38:36.000Z",
  "downloads": 48,
  "gated": false,
  "id": "selorahomes/Selora-AI-LLM-14B-ollama",
  "lastModified": "2026-10-05T05:13:44.000Z",
  "likes": 0,
  "modelId": "selorahomes/Selora-AI-LLM-14B-ollama",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "e8fcda46794e2b2510c676e4bc0efaf87159ee5d",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "Modelfile"
    },
    {
      "rfilename": "NOTICE"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "manifest.json"
    },
    {
      "rfilename": "params"
    },
    {
      "rfilename": "selora-ollama.Q5_K_S.gguf"
    },
    {
      "rfilename": "system"
    },
    {
      "rfilename": "template"
    }
  ],
  "tags": [
    "gguf",
    "text-generation",
    "qwen",
    "qwen3",
    "home-assistant",
    "home-automation",
    "smart-home",
    "ollama",
    "en",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Author

author

stefancosma

receipt

Source: GGUF quantisations
Its words: stefancosma
Read by: field:author
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aa662dd8ea16d584632df53",
  "author": "stefancosma",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "library_name": "transformers",
    "license": "apache-2.0",
    "license_link": "https://huggingface.co/Qwen/Qwen3-14B/blob/main/LICENSE",
    "pipeline_tag": "text-generation",
    "tags": [
      "llama-cpp",
      "gguf-my-repo"
    ]
  },
  "createdAt": "2026-09-13T08:46:21.000Z",
  "downloads": 186,
  "gated": false,
  "id": "stefancosma/Qwen3-14B-Q4_K_M-GGUF",
  "lastModified": "2026-09-13T08:46:59.000Z",
  "library_name": "transformers",
  "likes": 0,
  "modelId": "stefancosma/Qwen3-14B-Q4_K_M-GGUF",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "b13f355f7eb328c7093dee4ecb865e4ecb715f22",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "qwen3-14b-q4_k_m.gguf"
    }
  ],
  "tags": [
    "transformers",
    "gguf",
    "llama-cpp",
    "gguf-my-repo",
    "text-generation",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Author

author

stellatecnologia

receipt

Source: GGUF quantisations
Its words: stellatecnologia
Read by: field:author
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aa3137b9636871d471375d8",
  "author": "stellatecnologia",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "pt",
      "en",
      "fr"
    ],
    "library_name": "gguf",
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "santos-dumont",
      "historical-persona",
      "qwen3",
      "gguf",
      "stella-tecnologia"
    ]
  },
  "createdAt": "2026-09-10T20:30:51.000Z",
  "downloads": 10,
  "gated": false,
  "id": "stellatecnologia/santos-dumont-qwen3-14b",
  "lastModified": "2026-09-11T17:01:22.000Z",
  "library_name": "gguf",
  "likes": 1,
  "modelId": "stellatecnologia/santos-dumont-qwen3-14b",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "3da3c6a13d828c31e34e3d612a693876d602c5e8",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "CHANGELOG.md"
    },
    {
      "rfilename": "CITATION.cff"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "TECHNICAL_REPORT.md"
    },
    {
      "rfilename": "sd-v3.Q8_0.gguf"
    }
  ],
  "tags": [
    "gguf",
    "santos-dumont",
    "historical-persona",
    "qwen3",
    "stella-tecnologia",
    "text-generation",
    "pt",
    "en",
    "fr",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Base

base

GGUF quantisations

Qwen/Qwen3-14B

receipt

Source: GGUF quantisations
Its words: Qwen/Qwen3-14B
Read by: field:cardData.base_model[]
Said since: 2026-09-28 11:46 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6ab2aa765a05167560ef154e",
  "author": "Walkowicz",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en",
      "pt"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "gguf",
      "qwen3",
      "text-generation"
    ]
  },
  "createdAt": "2026-09-22T16:19:02.000Z",
  "downloads": 25,
  "gated": false,
  "id": "Walkowicz/kairos-sotos-1.0-14B",
  "lastModified": "2026-09-22T19:47:32.000Z",
  "likes": 0,
  "modelId": "Walkowicz/kairos-sotos-1.0-14B",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "d1c97e5d75be337e86ad69ce35cccb9c4660625a",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Modelfile"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "kairos-logo.png"
    },
    {
      "rfilename": "kairos-sotos-1.0-14B-Q6_K.gguf"
    }
  ],
  "tags": [
    "gguf",
    "qwen3",
    "text-generation",
    "en",
    "pt",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Base

base

Qwen/Qwen3-14B, Qwen/Qwen3.6-35B-A3B, Qwen/Qwen3.8-27B, Qwen/Qwen3-8B

receipt

Source: GGUF quantisations
Its words: Qwen/Qwen3-14B, Qwen/Qwen3.6-35B-A3B, Qwen/Qwen3.8-27B, Qwen/Qwen3-8B
Read by: field:cardData.base_model[]
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aabb0e05ed906a7bbc3ec17",
  "author": "Om22s",
  "cardData": {
    "base_model": [
      "Qwen/Qwen3-14B",
      "Qwen/Qwen3.6-35B-A3B",
      "Qwen/Qwen3.8-27B",
      "Qwen/Qwen3-8B"
    ],
    "base_model_relation": "adapter",
    "language": [
      "en"
    ],
    "library_name": "peft",
    "license": "apache-2.0",
    "model-index": [
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3-14b-rightsclean-attrv1)",
        "results": [
          {
            "dataset": {
              "name": "Alexandria four-book product fixture (768 paired rows, temperature 0, batch 25, default prompt, reasoning off)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 61.7)",
                "type": "accuracy",
                "value": 73.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "Alexandria four-book product fixture, reasoning low budget 1024, JSON schema",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 66.1)",
                "type": "accuracy",
                "value": 74.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), reasoning low budget 1024",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 68.9)",
                "type": "accuracy",
                "value": 75.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.9)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine held-out PDNC novels, 2,655 rows, IQ3_XXS, reasoning low, michel2_full (raw; 90.9 with the roster-line echo read as the app reads it)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.6)",
                "type": "accuracy",
                "value": 71.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q3_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.4)",
                "type": "accuracy",
                "value": 99.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.2)",
                "type": "accuracy",
                "value": 92.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.9)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine held-out PDNC novels, 2,655 rows, IQ3_XXS, reasoning low, michel2_full (raw; 90.9 with the roster-line echo read as the app reads it)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.6)",
                "type": "accuracy",
                "value": 71.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3-8b-rightsclean-window25-michel2v2)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning low, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 77.0)",
                "type": "accuracy",
                "value": 82.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-8B Q4_K_M, RX 9070 XT",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), Q4_K_M, reasoning low, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 76.6)",
                "type": "accuracy",
                "value": 82.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-8B Q4_K_M, RX 9070 XT",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2)",
        "results": [
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ1_M, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.7)",
                "type": "accuracy",
                "value": 92.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ2_XXS, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.9)",
                "type": "accuracy",
                "value": 91.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ3_XXS, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 92.5)",
                "type": "accuracy",
                "value": 92.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-Q4_K_XL, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 93.2)",
                "type": "accuracy",
                "value": 91.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      }
    ],
    "pipeline_tag": "text-generation",
    "tags": [
      "lora",
      "peft",
      "speaker-attribution",
      "audiobook",
      "qwen3",
      "rights-reviewed"
    ]
  },
  "createdAt": "2026-09-17T09:20:32.000Z",
  "downloads": 165,
  "gated": false,
  "id": "Om22s/alexandria-qwen3-attribution",
  "lastModified": "2026-09-29T07:14:41.000Z",
  "library_name": "peft",
  "likes": 0,
  "modelId": "Om22s/alexandria-qwen3-attribution",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "f54a4cd9599f4f496eae474cf9f6daee40b6af22",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/qwen3-14b-rightsclean-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/qwen3-14b-rightsclean-seed2-attrv2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/qwen3-14b-riqua3ep-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/qwen3-14b-riquax2-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/README.md"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/qwen3-8b-rightsclean-window25-michel2v2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/README.md"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/qwen3.6-35b-a3b-rightsclean-michel2v1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/README.md"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/qwen3.6-35b-a3b-rightsclean-window25-michel2v2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/README.md"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/qwen3.8-27b-rightsclean-michel2v1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/training_manifest.json"
    },
    {
      "rfilename": "default_prompts_attribute.txt"
    },
    {
      "rfilename": "examples/infer.py"
    },
    {
      "rfilename": "examples/serve_llamacpp.sh"
    },
    {
      "rfilename": "prompts/attrv1.txt"
    },
    {
      "rfilename": "prompts/attrv2.txt"
    },
    {
      "rfilename": "prompts/michel2v1.txt"
    },
    {
      "rfilename": "prompts/michel2v2.txt"
    }
  ],
  "tags": [
    "peft",
    "safetensors",
    "gguf",
    "lora",
    "speaker-attribution",
    "audiobook",
    "qwen3",
    "rights-reviewed",
    "text-generation",
    "en",
    "base_model:Qwen/Qwen3-14B",
    "base_model:adapter:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "model-index",
    "region:us"
  ]
}

—

Downloads

downloads

GGUF quantisations

receipt

Source: GGUF quantisations
Its words: 10
Read by: field:downloads
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aa3137b9636871d471375d8",
  "author": "stellatecnologia",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "pt",
      "en",
      "fr"
    ],
    "library_name": "gguf",
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "santos-dumont",
      "historical-persona",
      "qwen3",
      "gguf",
      "stella-tecnologia"
    ]
  },
  "createdAt": "2026-09-10T20:30:51.000Z",
  "downloads": 10,
  "gated": false,
  "id": "stellatecnologia/santos-dumont-qwen3-14b",
  "lastModified": "2026-09-11T17:01:22.000Z",
  "library_name": "gguf",
  "likes": 1,
  "modelId": "stellatecnologia/santos-dumont-qwen3-14b",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "3da3c6a13d828c31e34e3d612a693876d602c5e8",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "CHANGELOG.md"
    },
    {
      "rfilename": "CITATION.cff"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "TECHNICAL_REPORT.md"
    },
    {
      "rfilename": "sd-v3.Q8_0.gguf"
    }
  ],
  "tags": [
    "gguf",
    "santos-dumont",
    "historical-persona",
    "qwen3",
    "stella-tecnologia",
    "text-generation",
    "pt",
    "en",
    "fr",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

115

receipt

Source: GGUF quantisations
Its words: 115
Read by: field:downloads
Said since: 2026-10-05 12:24 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

2026-10-05 12:24 UTC	115
2026-10-04 12:16 UTC	114
2026-10-03 12:09 UTC	109
2026-10-02 12:01 UTC	100
2026-09-29 09:50 UTC	0

What the source handed over

{
  "_id": "6abacb4e87cf5463185425f5",
  "author": "KitsuneOff",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "license": "apache-2.0",
    "license_link": "https://huggingface.co/Qwen/Qwen3-14B-GGUF/blob/main/LICENSE",
    "pipeline_tag": "text-generation"
  },
  "createdAt": "2026-09-28T20:17:18.000Z",
  "downloads": 115,
  "gated": false,
  "id": "KitsuneOff/Qwen3-14B-GGUF",
  "lastModified": "2026-09-28T20:17:21.000Z",
  "likes": 0,
  "modelId": "KitsuneOff/Qwen3-14B-GGUF",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "218e5a18729725edbfd0e3e2355d98ad8684a7f4",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "Qwen3-14B-Q4_K_M.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Q5_0.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Q5_K_M.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Q6_K.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Q8_0.gguf"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "params"
    }
  ],
  "tags": [
    "gguf",
    "text-generation",
    "arxiv:2309.00071",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

126

receipt

Source: GGUF quantisations
Its words: 126
Read by: field:downloads
Said since: 2026-10-05 12:24 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

2026-10-05 12:24 UTC	126
2026-10-04 12:16 UTC	0

What the source handed over

{
  "_id": "6ac22ed17eb43e9e3e0fa3f4",
  "author": "Irfanuruchi",
  "cardData": {
    "base_model": [
      "Qwen/Qwen3-14B"
    ],
    "datasets": [
      "Irfanuruchi/buildeng-v8-3b"
    ],
    "language": [
      "en"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "qwen3",
      "qwen",
      "gguf",
      "llama-cpp",
      "building-engineering",
      "hvac",
      "engineering",
      "fine-tuned"
    ]
  },
  "createdAt": "2026-10-04T10:47:45.000Z",
  "downloads": 126,
  "gated": false,
  "id": "Irfanuruchi/Qwen3-14B-BuildEng-GGUF",
  "lastModified": "2026-10-04T12:55:26.000Z",
  "likes": 0,
  "modelId": "Irfanuruchi/Qwen3-14B-BuildEng-GGUF",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "6b29c423269d065b3974c5b91ce60c7065d3cafc",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "PUBLICATION-SHA256SUMS.txt"
    },
    {
      "rfilename": "Qwen3-14B-BuildEng-Q4_K_M.gguf"
    },
    {
      "rfilename": "Qwen3-14B-BuildEng-Q5_K_M.gguf"
    },
    {
      "rfilename": "Qwen3-14B-BuildEng-Q8_0.gguf"
    },
    {
      "rfilename": "README.md"
    }
  ],
  "tags": [
    "gguf",
    "qwen3",
    "qwen",
    "llama-cpp",
    "building-engineering",
    "hvac",
    "engineering",
    "fine-tuned",
    "text-generation",
    "en",
    "dataset:Irfanuruchi/buildeng-v8-3b",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

165

receipt

Source: GGUF quantisations
Its words: 165
Read by: field:downloads
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aabb0e05ed906a7bbc3ec17",
  "author": "Om22s",
  "cardData": {
    "base_model": [
      "Qwen/Qwen3-14B",
      "Qwen/Qwen3.6-35B-A3B",
      "Qwen/Qwen3.8-27B",
      "Qwen/Qwen3-8B"
    ],
    "base_model_relation": "adapter",
    "language": [
      "en"
    ],
    "library_name": "peft",
    "license": "apache-2.0",
    "model-index": [
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3-14b-rightsclean-attrv1)",
        "results": [
          {
            "dataset": {
              "name": "Alexandria four-book product fixture (768 paired rows, temperature 0, batch 25, default prompt, reasoning off)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 61.7)",
                "type": "accuracy",
                "value": 73.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "Alexandria four-book product fixture, reasoning low budget 1024, JSON schema",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 66.1)",
                "type": "accuracy",
                "value": 74.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), reasoning low budget 1024",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 68.9)",
                "type": "accuracy",
                "value": 75.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.9)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine held-out PDNC novels, 2,655 rows, IQ3_XXS, reasoning low, michel2_full (raw; 90.9 with the roster-line echo read as the app reads it)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.6)",
                "type": "accuracy",
                "value": 71.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.8-27b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.4)",
                "type": "accuracy",
                "value": 88.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.1)",
                "type": "accuracy",
                "value": 99.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 89.8)",
                "type": "accuracy",
                "value": 89.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.7)",
                "type": "accuracy",
                "value": 99.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q3_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 88.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q3_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 99.4)",
                "type": "accuracy",
                "value": 99.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B Q3_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 83.1)",
                "type": "accuracy",
                "value": 85.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.2)",
                "type": "accuracy",
                "value": 92.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.8-27B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-michel2v1)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning off, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 82.9)",
                "type": "accuracy",
                "value": 85.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning off, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 97.8)",
                "type": "accuracy",
                "value": 96.2,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ1_M, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 85,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ1_M, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.2)",
                "type": "accuracy",
                "value": 98.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ2_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 88.3)",
                "type": "accuracy",
                "value": 87.9,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ2_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, IQ3_XXS, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 87.5)",
                "type": "accuracy",
                "value": 86.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), IQ3_XXS, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 95.6)",
                "type": "accuracy",
                "value": 82.7,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_XL, reasoning on, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.8)",
                "type": "accuracy",
                "value": 88.5,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "PDNC Emma, held out (318 rows, 40 windows), Q4_K_XL, reasoning on, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 96.9)",
                "type": "accuracy",
                "value": 95.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine held-out PDNC novels, 2,655 rows, IQ3_XXS, reasoning low, michel2_full (raw; 90.9 with the roster-line echo read as the app reads it)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 91.6)",
                "type": "accuracy",
                "value": 71.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA, Qwen3.6-35B-A3B IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3-8b-rightsclean-window25-michel2v2)",
        "results": [
          {
            "dataset": {
              "name": "four-book product fixture, Q4_K_M, reasoning low, michel2_full, JSON schema, 768 paired rows",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 77.0)",
                "type": "accuracy",
                "value": 82.3,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-8B Q4_K_M, RX 9070 XT",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), Q4_K_M, reasoning low, michel2_full",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 76.6)",
                "type": "accuracy",
                "value": 82.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-8B Q4_K_M, RX 9070 XT",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      },
      {
        "name": "alexandria-qwen3-attribution (adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2)",
        "results": [
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ1_M, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.7)",
                "type": "accuracy",
                "value": 92.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ1_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ2_XXS, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 90.9)",
                "type": "accuracy",
                "value": 91.6,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ2_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-IQ3_XXS, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 92.5)",
                "type": "accuracy",
                "value": 92.8,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-IQ3_XXS",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          },
          {
            "dataset": {
              "name": "nine PDNC novels, all held out (2,655 rows, 40 windows each), UD-Q4_K_XL, reasoning low, michel2_full (michel2v2)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 93.2)",
                "type": "accuracy",
                "value": 91.1,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3.6-35B-A3B UD-Q4_K_XL",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      }
    ],
    "pipeline_tag": "text-generation",
    "tags": [
      "lora",
      "peft",
      "speaker-attribution",
      "audiobook",
      "qwen3",
      "rights-reviewed"
    ]
  },
  "createdAt": "2026-09-17T09:20:32.000Z",
  "downloads": 165,
  "gated": false,
  "id": "Om22s/alexandria-qwen3-attribution",
  "lastModified": "2026-09-29T07:14:41.000Z",
  "library_name": "peft",
  "likes": 0,
  "modelId": "Om22s/alexandria-qwen3-attribution",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "f54a4cd9599f4f496eae474cf9f6daee40b6af22",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/qwen3-14b-rightsclean-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/qwen3-14b-rightsclean-seed2-attrv2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-rightsclean-seed2-attrv2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/qwen3-14b-riqua3ep-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riqua3ep-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/README.md"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/qwen3-14b-riquax2-attrv1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-14b-riquax2-attrv1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/README.md"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/qwen3-8b-rightsclean-window25-michel2v2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3-8b-rightsclean-window25-michel2v2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/README.md"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/qwen3.6-35b-a3b-rightsclean-michel2v1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-michel2v1/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/README.md"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/qwen3.6-35b-a3b-rightsclean-window25-michel2v2.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.6-35b-a3b-rightsclean-window25-michel2v2/training_manifest.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/README.md"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/adapter_config.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/adapter_model.safetensors"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/qwen3.8-27b-rightsclean-michel2v1.f16.gguf"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/tokenizer_config.json"
    },
    {
      "rfilename": "adapters/qwen3.8-27b-rightsclean-michel2v1/training_manifest.json"
    },
    {
      "rfilename": "default_prompts_attribute.txt"
    },
    {
      "rfilename": "examples/infer.py"
    },
    {
      "rfilename": "examples/serve_llamacpp.sh"
    },
    {
      "rfilename": "prompts/attrv1.txt"
    },
    {
      "rfilename": "prompts/attrv2.txt"
    },
    {
      "rfilename": "prompts/michel2v1.txt"
    },
    {
      "rfilename": "prompts/michel2v2.txt"
    }
  ],
  "tags": [
    "peft",
    "safetensors",
    "gguf",
    "lora",
    "speaker-attribution",
    "audiobook",
    "qwen3",
    "rights-reviewed",
    "text-generation",
    "en",
    "base_model:Qwen/Qwen3-14B",
    "base_model:adapter:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "model-index",
    "region:us"
  ]
}

—

Downloads

downloads

186

receipt

Source: GGUF quantisations
Its words: 186
Read by: field:downloads
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aa662dd8ea16d584632df53",
  "author": "stefancosma",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "library_name": "transformers",
    "license": "apache-2.0",
    "license_link": "https://huggingface.co/Qwen/Qwen3-14B/blob/main/LICENSE",
    "pipeline_tag": "text-generation",
    "tags": [
      "llama-cpp",
      "gguf-my-repo"
    ]
  },
  "createdAt": "2026-09-13T08:46:21.000Z",
  "downloads": 186,
  "gated": false,
  "id": "stefancosma/Qwen3-14B-Q4_K_M-GGUF",
  "lastModified": "2026-09-13T08:46:59.000Z",
  "library_name": "transformers",
  "likes": 0,
  "modelId": "stefancosma/Qwen3-14B-Q4_K_M-GGUF",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "b13f355f7eb328c7093dee4ecb865e4ecb715f22",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "qwen3-14b-q4_k_m.gguf"
    }
  ],
  "tags": [
    "transformers",
    "gguf",
    "llama-cpp",
    "gguf-my-repo",
    "text-generation",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

receipt

Source: GGUF quantisations
Its words: 20
Read by: field:downloads
Said since: 2026-10-05 12:24 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

2026-10-05 12:24 UTC	20
2026-10-05 06:21 UTC	0

What the source handed over

{
  "_id": "6ac3105fea36a8977d078842",
  "author": "anyalkonsafe",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en"
    ],
    "library_name": "peft",
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "qwen",
      "qwen3",
      "gguf",
      "lora",
      "qlora",
      "unsloth",
      "persona",
      "tool-calling",
      "text-generation"
    ]
  },
  "createdAt": "2026-10-05T02:50:07.000Z",
  "downloads": 20,
  "gated": false,
  "id": "anyalkonsafe/evilguy-3-gguf",
  "lastModified": "2026-10-05T02:55:33.000Z",
  "library_name": "peft",
  "likes": 0,
  "modelId": "anyalkonsafe/evilguy-3-gguf",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "f130250fe6b4141bd51dc886512ce57157068643",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "evilguy-3-q4_k_m.gguf"
    }
  ],
  "tags": [
    "peft",
    "gguf",
    "qwen",
    "qwen3",
    "lora",
    "qlora",
    "unsloth",
    "persona",
    "tool-calling",
    "text-generation",
    "en",
    "base_model:Qwen/Qwen3-14B",
    "base_model:adapter:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

248

receipt

Source: GGUF quantisations
Its words: 248
Read by: field:downloads
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6a9b8f6727087a68413ae2cc",
  "author": "chenhaodev",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "zh"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "medical",
      "internal-medicine",
      "qwen",
      "gguf",
      "llama.cpp",
      "chinese"
    ]
  },
  "createdAt": "2026-09-05T03:41:27.000Z",
  "downloads": 248,
  "gated": false,
  "id": "chenhaodev/qwen3-14b-internal-med-gguf",
  "lastModified": "2026-09-05T05:00:01.000Z",
  "likes": 0,
  "modelId": "chenhaodev/qwen3-14b-internal-med-gguf",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "cc9ac7828cc7c901b54673a4d73e9b7d64425779",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Qwen3-14B-Internal-Med-q4_k_m.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Internal-Med-q5_k_m.gguf"
    },
    {
      "rfilename": "Qwen3-14B-Internal-Med-q8_0.gguf"
    },
    {
      "rfilename": "README.md"
    }
  ],
  "tags": [
    "gguf",
    "medical",
    "internal-medicine",
    "qwen",
    "llama.cpp",
    "chinese",
    "text-generation",
    "zh",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

receipt

Source: GGUF quantisations
Its words: 25
Read by: field:downloads
Said since: 2026-09-28 11:46 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6ab2aa765a05167560ef154e",
  "author": "Walkowicz",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en",
      "pt"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "gguf",
      "qwen3",
      "text-generation"
    ]
  },
  "createdAt": "2026-09-22T16:19:02.000Z",
  "downloads": 25,
  "gated": false,
  "id": "Walkowicz/kairos-sotos-1.0-14B",
  "lastModified": "2026-09-22T19:47:32.000Z",
  "likes": 0,
  "modelId": "Walkowicz/kairos-sotos-1.0-14B",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "d1c97e5d75be337e86ad69ce35cccb9c4660625a",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Modelfile"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "kairos-logo.png"
    },
    {
      "rfilename": "kairos-sotos-1.0-14B-Q6_K.gguf"
    }
  ],
  "tags": [
    "gguf",
    "qwen3",
    "text-generation",
    "en",
    "pt",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

receipt

Source: GGUF quantisations
Its words: 48
Read by: field:downloads
Said since: 2026-10-05 12:24 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

2026-10-05 12:24 UTC	48
2026-10-04 12:16 UTC	47
2026-10-03 12:09 UTC	34
2026-10-02 12:01 UTC	26

What the source handed over

{
  "_id": "6abbf79c51f0022e9a4225f9",
  "author": "selorahomes",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "text-generation",
      "qwen",
      "qwen3",
      "gguf",
      "home-assistant",
      "home-automation",
      "smart-home",
      "ollama"
    ]
  },
  "createdAt": "2026-09-29T17:38:36.000Z",
  "downloads": 48,
  "gated": false,
  "id": "selorahomes/Selora-AI-LLM-14B-ollama",
  "lastModified": "2026-10-05T05:13:44.000Z",
  "likes": 0,
  "modelId": "selorahomes/Selora-AI-LLM-14B-ollama",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "e8fcda46794e2b2510c676e4bc0efaf87159ee5d",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "Modelfile"
    },
    {
      "rfilename": "NOTICE"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "manifest.json"
    },
    {
      "rfilename": "params"
    },
    {
      "rfilename": "selora-ollama.Q5_K_S.gguf"
    },
    {
      "rfilename": "system"
    },
    {
      "rfilename": "template"
    }
  ],
  "tags": [
    "gguf",
    "text-generation",
    "qwen",
    "qwen3",
    "home-assistant",
    "home-automation",
    "smart-home",
    "ollama",
    "en",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

receipt

Source: GGUF quantisations
Its words: 52
Read by: field:downloads
Said since: 2026-10-02 12:01 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

2026-10-02 12:01 UTC	52
2026-09-28 11:46 UTC	46

What the source handed over

{
  "_id": "6ab7c98640a9cf563795b362",
  "author": "1bit-MONSTER",
  "cardData": {
    "base_model": [
      "Qwen/Qwen3-14B"
    ],
    "license": "apache-2.0",
    "tags": [
      "gguf"
    ]
  },
  "createdAt": "2026-09-26T13:32:54.000Z",
  "downloads": 52,
  "gated": false,
  "id": "1bit-MONSTER/Qwen3-14B-GGUF",
  "lastModified": "2026-09-26T13:33:22.000Z",
  "likes": 0,
  "modelId": "1bit-MONSTER/Qwen3-14B-GGUF",
  "private": false,
  "sha": "35bd73c5af9de2073e03339e86c059c236519abb",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Qwen3-14B-Q4_K_M.gguf"
    },
    {
      "rfilename": "README.md"
    }
  ],
  "tags": [
    "gguf",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Downloads

downloads

receipt

Source: GGUF quantisations
Its words: 85
Read by: field:downloads
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aa9aa25d2fcda3a1c4fb8f9",
  "author": "Om22s",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "base_model_relation": "adapter",
    "language": [
      "en"
    ],
    "library_name": "peft",
    "license": "apache-2.0",
    "model-index": [
      {
        "name": "alexandria-qwen3-14b-rightsclean-speaker-attribution",
        "results": [
          {
            "dataset": {
              "name": "Alexandria four-book product fixture (768 paired rows, temperature 0, batch 25, default prompt, reasoning off)",
              "type": "alexandria-attribution-fixture"
            },
            "metrics": [
              {
                "name": "accuracy, adapter (base 61.7)",
                "type": "accuracy",
                "value": 73.4,
                "verified": false
              }
            ],
            "source": {
              "name": "paired base/LoRA on one llama.cpp server, Qwen3-14B Q4_K_M",
              "url": "https://github.com/on22s/alexandria-audiobook2/blob/main/RECIPES.md"
            },
            "task": {
              "name": "speaker attribution (structured JSON, roster given)",
              "type": "text-generation"
            }
          }
        ]
      }
    ],
    "new_version": "Om22s/alexandria-qwen3-attribution",
    "pipeline_tag": "text-generation",
    "tags": [
      "peft",
      "lora",
      "speaker-attribution",
      "audiobook",
      "rights-reviewed"
    ]
  },
  "createdAt": "2026-09-15T20:27:17.000Z",
  "downloads": 85,
  "gated": false,
  "id": "Om22s/alexandria-qwen3-14b-rightsclean-speaker-attribution-attrv1",
  "lastModified": "2026-09-27T21:11:07.000Z",
  "library_name": "peft",
  "likes": 0,
  "modelId": "Om22s/alexandria-qwen3-14b-rightsclean-speaker-attribution-attrv1",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "f896ea59f8f97712c633c37566bf3302fce67571",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "LICENSE"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "adapter_config.json"
    },
    {
      "rfilename": "adapter_model.safetensors"
    },
    {
      "rfilename": "default_prompts_attribute.txt"
    },
    {
      "rfilename": "examples/infer.py"
    },
    {
      "rfilename": "examples/serve_llamacpp.sh"
    },
    {
      "rfilename": "rightsclean-attrv1.f16.gguf"
    },
    {
      "rfilename": "tokenizer_config.json"
    }
  ],
  "tags": [
    "peft",
    "safetensors",
    "gguf",
    "lora",
    "speaker-attribution",
    "audiobook",
    "rights-reviewed",
    "text-generation",
    "en",
    "base_model:Qwen/Qwen3-14B",
    "base_model:adapter:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "model-index",
    "region:us"
  ]
}

—

Licence

licence

GGUF quantisations

apache-2.0

receipt

Source: GGUF quantisations
Its words: apache-2.0
Read by: field:cardData.license
Said since: 2026-09-28 11:46 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6ab2aa765a05167560ef154e",
  "author": "Walkowicz",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en",
      "pt"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "gguf",
      "qwen3",
      "text-generation"
    ]
  },
  "createdAt": "2026-09-22T16:19:02.000Z",
  "downloads": 25,
  "gated": false,
  "id": "Walkowicz/kairos-sotos-1.0-14B",
  "lastModified": "2026-09-22T19:47:32.000Z",
  "likes": 0,
  "modelId": "Walkowicz/kairos-sotos-1.0-14B",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "d1c97e5d75be337e86ad69ce35cccb9c4660625a",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Modelfile"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "kairos-logo.png"
    },
    {
      "rfilename": "kairos-sotos-1.0-14B-Q6_K.gguf"
    }
  ],
  "tags": [
    "gguf",
    "qwen3",
    "text-generation",
    "en",
    "pt",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Likes

likes

GGUF quantisations

receipt

Source: GGUF quantisations
Its words: 0
Read by: field:likes
Said since: 2026-09-28 11:46 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6ab2aa765a05167560ef154e",
  "author": "Walkowicz",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "en",
      "pt"
    ],
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "gguf",
      "qwen3",
      "text-generation"
    ]
  },
  "createdAt": "2026-09-22T16:19:02.000Z",
  "downloads": 25,
  "gated": false,
  "id": "Walkowicz/kairos-sotos-1.0-14B",
  "lastModified": "2026-09-22T19:47:32.000Z",
  "likes": 0,
  "modelId": "Walkowicz/kairos-sotos-1.0-14B",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "d1c97e5d75be337e86ad69ce35cccb9c4660625a",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "Modelfile"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "kairos-logo.png"
    },
    {
      "rfilename": "kairos-sotos-1.0-14B-Q6_K.gguf"
    }
  ],
  "tags": [
    "gguf",
    "qwen3",
    "text-generation",
    "en",
    "pt",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

Likes

likes

receipt

Source: GGUF quantisations
Its words: 1
Read by: field:likes
Said since: 2026-09-29 09:50 UTC
Last answered: 2026-10-05 12:24 UTC
Original: open at the source

What the source handed over

{
  "_id": "6aa3137b9636871d471375d8",
  "author": "stellatecnologia",
  "cardData": {
    "base_model": "Qwen/Qwen3-14B",
    "language": [
      "pt",
      "en",
      "fr"
    ],
    "library_name": "gguf",
    "license": "apache-2.0",
    "pipeline_tag": "text-generation",
    "tags": [
      "santos-dumont",
      "historical-persona",
      "qwen3",
      "gguf",
      "stella-tecnologia"
    ]
  },
  "createdAt": "2026-09-10T20:30:51.000Z",
  "downloads": 10,
  "gated": false,
  "id": "stellatecnologia/santos-dumont-qwen3-14b",
  "lastModified": "2026-09-11T17:01:22.000Z",
  "library_name": "gguf",
  "likes": 1,
  "modelId": "stellatecnologia/santos-dumont-qwen3-14b",
  "pipeline_tag": "text-generation",
  "private": false,
  "sha": "3da3c6a13d828c31e34e3d612a693876d602c5e8",
  "siblings": [
    {
      "rfilename": ".gitattributes"
    },
    {
      "rfilename": "CHANGELOG.md"
    },
    {
      "rfilename": "CITATION.cff"
    },
    {
      "rfilename": "README.md"
    },
    {
      "rfilename": "TECHNICAL_REPORT.md"
    },
    {
      "rfilename": "sd-v3.Q8_0.gguf"
    }
  ],
  "tags": [
    "gguf",
    "santos-dumont",
    "historical-persona",
    "qwen3",
    "stella-tecnologia",
    "text-generation",
    "pt",
    "en",
    "fr",
    "base_model:Qwen/Qwen3-14B",
    "base_model:quantized:Qwen/Qwen3-14B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ]
}

—

quantisation

Walkowicz/kairos-sotos-1.0-14B zetlyn/models-gguf · 2026-09-22	author Walkowicz base Qwen/Qwen3-14B downloads 25 licence apache-2.0 likes 0	source
Om22s/alexandria-qwen3-attribution zetlyn/models-gguf · 2026-09-17	author Om22s base Qwen/Qwen3-14B, Qwen/Qwen3.6-35B-A3B, Qwen/Qwen3.8-27B, Qwen/Qwen3-8B downloads 165 licence apache-2.0 likes 0	source
Om22s/alexandria-qwen3-14b-rightsclean-speaker-attribution-attrv1 zetlyn/models-gguf · 2026-09-15	author Om22s base Qwen/Qwen3-14B downloads 85 licence apache-2.0 likes 0	source
stefancosma/Qwen3-14B-Q4_K_M-GGUF zetlyn/models-gguf · 2026-09-13	author stefancosma base Qwen/Qwen3-14B downloads 186 licence apache-2.0 likes 0	source
stellatecnologia/santos-dumont-qwen3-14b zetlyn/models-gguf · 2026-09-10	author stellatecnologia base Qwen/Qwen3-14B downloads 10 licence apache-2.0 likes 1	source
chenhaodev/qwen3-14b-internal-med-gguf zetlyn/models-gguf · 2026-09-05	author chenhaodev base Qwen/Qwen3-14B downloads 248 licence apache-2.0 likes 0	source
1bit-MONSTER/Qwen3-14B-GGUF zetlyn/models-gguf · 2026-09-26	author 1bit-MONSTER base Qwen/Qwen3-14B downloads 52 licence apache-2.0 likes 0	source
anyalkonsafe/evilguy-3-gguf zetlyn/models-gguf · 2026-10-05	author anyalkonsafe base Qwen/Qwen3-14B downloads 20 licence apache-2.0 likes 0	source
Irfanuruchi/Qwen3-14B-BuildEng-GGUF zetlyn/models-gguf · 2026-10-04	author Irfanuruchi base Qwen/Qwen3-14B downloads 126 licence apache-2.0 likes 0	source
selorahomes/Selora-AI-LLM-14B-ollama zetlyn/models-gguf · 2026-09-29	author selorahomes base Qwen/Qwen3-14B downloads 48 licence apache-2.0 likes 0	source
KitsuneOff/Qwen3-14B-GGUF zetlyn/models-gguf · 2026-09-28	author KitsuneOff base Qwen/Qwen3-14B downloads 115 licence apache-2.0 likes 0	source