{
  "created_at": "2026-05-19T18:39:45.207640+00:00",
  "config": {
    "hf_dataset": "faodl/amis-agri-production-v03",
    "hf_subset": null,
    "train_split": "train",
    "validation_split": "validation",
    "test_split": "test",
    "text_col": "chunk_text",
    "label_col": "label",
    "group_col": "id",
    "id_col": "chunk_id",
    "model_name": "FacebookAI/xlm-roberta-base",
    "output_dir": "/content/agri-production-classifier",
    "max_length": 256,
    "learning_rate": 2e-05,
    "weight_decay": 0.01,
    "num_train_epochs": 5.0,
    "per_device_train_batch_size": 16,
    "per_device_eval_batch_size": 32,
    "gradient_accumulation_steps": 1,
    "warmup_ratio": 0.1,
    "early_stopping_patience": 2,
    "seed": 42,
    "metric_for_best_model": "f1",
    "skip_transformer": false,
    "skip_baselines": false,
    "baseline_models": [
      "logistic",
      "xgboost",
      "embedding-logistic",
      "embedding-svm",
      "embedding-lightgbm"
    ],
    "tfidf_max_features": 50000,
    "embedding_model_name": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
    "embedding_batch_size": 64,
    "positive_label_name": "RELEVANT",
    "negative_label_name": "NOT_RELEVANT",
    "push_to_hub": true,
    "hub_model_id": "faodl/agri-production-classifier",
    "hub_private_repo": false
  },
  "dataset_summary": {
    "train": {
      "rows": 5937,
      "labels": {
        "0": 4830,
        "1": 1107
      },
      "unique_groups": null,
      "text_length_mean": 670.8157318511032,
      "text_length_median": 526.0
    },
    "validation": {
      "rows": 1280,
      "labels": {
        "0": 1011,
        "1": 269
      },
      "unique_groups": null,
      "text_length_mean": 678.05,
      "text_length_median": 513.0
    },
    "test": {
      "rows": 1388,
      "labels": {
        "0": 1106,
        "1": 282
      },
      "unique_groups": null,
      "text_length_mean": 659.2327089337176,
      "text_length_median": 517.5
    }
  },
  "results": [
    {
      "model_type": "logistic_tfidf",
      "model_name": "logistic",
      "artifact_dir": "/content/agri-production-classifier/baselines/logistic",
      "artifact_file": "/content/agri-production-classifier/baselines/logistic/logistic_tfidf.joblib",
      "validation_best_threshold": {
        "threshold": 0.47902895736910023,
        "f1": 0.7775831873905429,
        "precision": 0.7350993377483444,
        "recall": 0.8252788104089219
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.899135446685879,
        "precision": 0.7414965986394558,
        "recall": 0.7730496453900709,
        "f1": 0.7569444444444444,
        "confusion_matrix": [
          [
            1030,
            76
          ],
          [
            64,
            218
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9414990859232175,
            "recall": 0.9312839059674503,
            "f1-score": 0.9363636363636364,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.7414965986394558,
            "recall": 0.7730496453900709,
            "f1-score": 0.7569444444444444,
            "support": 282.0
          },
          "accuracy": 0.899135446685879,
          "macro avg": {
            "precision": 0.8414978422813366,
            "recall": 0.8521667756787605,
            "f1-score": 0.8466540404040404,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.9008645748180151,
            "recall": 0.899135446685879,
            "f1-score": 0.8999110339708323,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9343298321213753,
        "average_precision": 0.8131443889860821
      },
      "test_optimal_threshold": {
        "threshold": 0.47902895736910023,
        "accuracy": 0.8969740634005764,
        "precision": 0.7293729372937293,
        "recall": 0.7836879432624113,
        "f1": 0.7555555555555555,
        "confusion_matrix": [
          [
            1024,
            82
          ],
          [
            61,
            221
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.943778801843318,
            "recall": 0.9258589511754068,
            "f1-score": 0.9347329986307622,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.7293729372937293,
            "recall": 0.7836879432624113,
            "f1-score": 0.7555555555555555,
            "support": 282.0
          },
          "accuracy": 0.8969740634005764,
          "macro avg": {
            "precision": 0.8365758695685237,
            "recall": 0.854773447218909,
            "f1-score": 0.8451442770931589,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.9002179561639346,
            "recall": 0.8969740634005764,
            "f1-score": 0.8983295123575573,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9343298321213753,
        "average_precision": 0.8131443889860821
      }
    },
    {
      "model_type": "xgboost_tfidf",
      "model_name": "xgboost",
      "artifact_dir": "/content/agri-production-classifier/baselines/xgboost",
      "artifact_file": "/content/agri-production-classifier/baselines/xgboost/xgboost_tfidf.joblib",
      "validation_best_threshold": {
        "threshold": 0.19665858149528503,
        "f1": 0.7173144876325089,
        "precision": 0.6835016835016835,
        "recall": 0.7546468401486989
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.8976945244956772,
        "precision": 0.8723404255319149,
        "recall": 0.5815602836879432,
        "f1": 0.6978723404255319,
        "confusion_matrix": [
          [
            1082,
            24
          ],
          [
            118,
            164
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9016666666666666,
            "recall": 0.9783001808318263,
            "f1-score": 0.9384215091066782,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.8723404255319149,
            "recall": 0.5815602836879432,
            "f1-score": 0.6978723404255319,
            "support": 282.0
          },
          "accuracy": 0.8976945244956772,
          "macro avg": {
            "precision": 0.8870035460992908,
            "recall": 0.7799302322598848,
            "f1-score": 0.8181469247661051,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.8957084534101826,
            "recall": 0.8976945244956772,
            "f1-score": 0.889549127573477,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9163893270747566,
        "average_precision": 0.7909075826875256
      },
      "test_optimal_threshold": {
        "threshold": 0.19665858149528503,
        "accuracy": 0.8782420749279539,
        "precision": 0.6738461538461539,
        "recall": 0.776595744680851,
        "f1": 0.7215815485996705,
        "confusion_matrix": [
          [
            1000,
            106
          ],
          [
            63,
            219
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.940733772342427,
            "recall": 0.9041591320072333,
            "f1-score": 0.9220839096357769,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.6738461538461539,
            "recall": 0.776595744680851,
            "f1-score": 0.7215815485996705,
            "support": 282.0
          },
          "accuracy": 0.8782420749279539,
          "macro avg": {
            "precision": 0.8072899630942905,
            "recall": 0.8403774383440421,
            "f1-score": 0.8218327291177236,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.8865102072012533,
            "recall": 0.8782420749279539,
            "f1-score": 0.8813478391659052,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9163893270747566,
        "average_precision": 0.7909075826875256
      }
    },
    {
      "model_type": "embedding-logistic_sentence_embeddings",
      "model_name": "logistic",
      "embedding_model_name": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
      "artifact_dir": "/content/agri-production-classifier/baselines/embedding-logistic",
      "artifact_file": "/content/agri-production-classifier/baselines/embedding-logistic/embedding-logistic.joblib",
      "validation_best_threshold": {
        "threshold": 0.504836451018676,
        "f1": 0.7974481658692185,
        "precision": 0.6983240223463687,
        "recall": 0.929368029739777
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.8796829971181557,
        "precision": 0.6455696202531646,
        "recall": 0.9042553191489362,
        "f1": 0.7533234859675036,
        "confusion_matrix": [
          [
            966,
            140
          ],
          [
            27,
            255
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.972809667673716,
            "recall": 0.8734177215189873,
            "f1-score": 0.920438303954264,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.6455696202531646,
            "recall": 0.9042553191489362,
            "f1-score": 0.7533234859675036,
            "support": 282.0
          },
          "accuracy": 0.8796829971181557,
          "macro avg": {
            "precision": 0.8091896439634403,
            "recall": 0.8888365203339618,
            "f1-score": 0.8368808949608838,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.9063242978087336,
            "recall": 0.8796829971181557,
            "f1-score": 0.8864855815679049,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9491618893719621,
        "average_precision": 0.8123161283338098
      },
      "test_optimal_threshold": {
        "threshold": 0.504836451018676,
        "accuracy": 0.8804034582132565,
        "precision": 0.6472081218274112,
        "recall": 0.9042553191489362,
        "f1": 0.7544378698224852,
        "confusion_matrix": [
          [
            967,
            139
          ],
          [
            27,
            255
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9728370221327968,
            "recall": 0.8743218806509946,
            "f1-score": 0.920952380952381,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.6472081218274112,
            "recall": 0.9042553191489362,
            "f1-score": 0.7544378698224852,
            "support": 282.0
          },
          "accuracy": 0.8804034582132565,
          "macro avg": {
            "precision": 0.810022571980104,
            "recall": 0.8892885998999653,
            "f1-score": 0.8376951253874331,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.9066789890736334,
            "recall": 0.8804034582132565,
            "f1-score": 0.887121622927431,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9491618893719621,
        "average_precision": 0.8123161283338098
      }
    },
    {
      "model_type": "embedding-svm_sentence_embeddings",
      "model_name": "svm",
      "embedding_model_name": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
      "artifact_dir": "/content/agri-production-classifier/baselines/embedding-svm",
      "artifact_file": "/content/agri-production-classifier/baselines/embedding-svm/embedding-svm.joblib",
      "validation_best_threshold": {
        "threshold": 0.26135853709254125,
        "f1": 0.7848932676518883,
        "precision": 0.7029411764705882,
        "recall": 0.8884758364312267
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.8847262247838616,
        "precision": 0.7276119402985075,
        "recall": 0.6914893617021277,
        "f1": 0.7090909090909091,
        "confusion_matrix": [
          [
            1033,
            73
          ],
          [
            87,
            195
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9223214285714286,
            "recall": 0.933996383363472,
            "f1-score": 0.9281221922731356,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.7276119402985075,
            "recall": 0.6914893617021277,
            "f1-score": 0.7090909090909091,
            "support": 282.0
          },
          "accuracy": 0.8847262247838616,
          "macro avg": {
            "precision": 0.8249666844349681,
            "recall": 0.8127428725327999,
            "f1-score": 0.8186065506820224,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.8827622962277948,
            "recall": 0.8847262247838616,
            "f1-score": 0.8836216001568619,
            "support": 1388.0
          }
        },
        "roc_auc": 0.946846985494979,
        "average_precision": 0.8073861124441263
      },
      "test_optimal_threshold": {
        "threshold": 0.26135853709254125,
        "accuracy": 0.8832853025936599,
        "precision": 0.6630434782608695,
        "recall": 0.8652482269503546,
        "f1": 0.7507692307692307,
        "confusion_matrix": [
          [
            982,
            124
          ],
          [
            38,
            244
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9627450980392157,
            "recall": 0.8878842676311031,
            "f1-score": 0.9238005644402634,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.6630434782608695,
            "recall": 0.8652482269503546,
            "f1-score": 0.7507692307692307,
            "support": 282.0
          },
          "accuracy": 0.8832853025936599,
          "macro avg": {
            "precision": 0.8128942881500426,
            "recall": 0.8765662472907289,
            "f1-score": 0.8372848976047471,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.9018547113119149,
            "recall": 0.8832853025936599,
            "f1-score": 0.888645783391826,
            "support": 1388.0
          }
        },
        "roc_auc": 0.946846985494979,
        "average_precision": 0.8073861124441263
      }
    },
    {
      "model_type": "embedding-lightgbm_sentence_embeddings",
      "model_name": "lightgbm",
      "embedding_model_name": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
      "artifact_dir": "/content/agri-production-classifier/baselines/embedding-lightgbm",
      "artifact_file": "/content/agri-production-classifier/baselines/embedding-lightgbm/embedding-lightgbm.joblib",
      "validation_best_threshold": {
        "threshold": 0.16106461000649094,
        "f1": 0.8150684931506849,
        "precision": 0.7555555555555555,
        "recall": 0.8847583643122676
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.8904899135446686,
        "precision": 0.7407407407407407,
        "recall": 0.7092198581560284,
        "f1": 0.7246376811594203,
        "confusion_matrix": [
          [
            1036,
            70
          ],
          [
            82,
            200
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9266547406082289,
            "recall": 0.9367088607594937,
            "f1-score": 0.9316546762589928,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.7407407407407407,
            "recall": 0.7092198581560284,
            "f1-score": 0.7246376811594203,
            "support": 282.0
          },
          "accuracy": 0.8904899135446686,
          "macro avg": {
            "precision": 0.8336977406744848,
            "recall": 0.822964359457761,
            "f1-score": 0.8281461787092066,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.8888825878973993,
            "recall": 0.8904899135446686,
            "f1-score": 0.8895950273987049,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9495209880343196,
        "average_precision": 0.8095411157881081
      },
      "test_optimal_threshold": {
        "threshold": 0.16106461000649094,
        "accuracy": 0.8955331412103746,
        "precision": 0.7044776119402985,
        "recall": 0.8368794326241135,
        "f1": 0.7649918962722853,
        "confusion_matrix": [
          [
            1007,
            99
          ],
          [
            46,
            236
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.956315289648623,
            "recall": 0.9104882459312839,
            "f1-score": 0.9328392774432608,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.7044776119402985,
            "recall": 0.8368794326241135,
            "f1-score": 0.7649918962722853,
            "support": 282.0
          },
          "accuracy": 0.8955331412103746,
          "macro avg": {
            "precision": 0.8303964507944608,
            "recall": 0.8736838392776987,
            "f1-score": 0.848915586857773,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.9051494214110528,
            "recall": 0.8955331412103746,
            "f1-score": 0.8987377201736533,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9495209880343196,
        "average_precision": 0.8095411157881081
      }
    },
    {
      "model_type": "transformer",
      "model_name": "FacebookAI/xlm-roberta-base",
      "artifact_dir": "/content/agri-production-classifier/transformer",
      "validation_best_threshold": {
        "threshold": 0.3764225244522095,
        "f1": 0.8217317487266553,
        "precision": 0.75625,
        "recall": 0.8996282527881041
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.9012968299711815,
        "precision": 0.7331189710610932,
        "recall": 0.8085106382978723,
        "f1": 0.7689713322091062,
        "confusion_matrix": [
          [
            1023,
            83
          ],
          [
            54,
            228
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9498607242339833,
            "recall": 0.9249547920433996,
            "f1-score": 0.9372423270728355,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.7331189710610932,
            "recall": 0.8085106382978723,
            "f1-score": 0.7689713322091062,
            "support": 282.0
          },
          "accuracy": 0.9012968299711815,
          "macro avg": {
            "precision": 0.8414898476475383,
            "recall": 0.8667327151706359,
            "f1-score": 0.8531068296409708,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.9058252959956873,
            "recall": 0.9012968299711815,
            "f1-score": 0.9030547041970635,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9566660895438165,
        "average_precision": 0.8271081453687152
      },
      "test_optimal_threshold": {
        "threshold": 0.3764225244522095,
        "accuracy": 0.8962536023054755,
        "precision": 0.6960227272727273,
        "recall": 0.8687943262411347,
        "f1": 0.7728706624605678,
        "confusion_matrix": [
          [
            999,
            107
          ],
          [
            37,
            245
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9642857142857143,
            "recall": 0.903254972875226,
            "f1-score": 0.9327731092436975,
            "support": 1106.0
          },
          "RELEVANT": {
            "precision": 0.6960227272727273,
            "recall": 0.8687943262411347,
            "f1-score": 0.7728706624605678,
            "support": 282.0
          },
          "accuracy": 0.8962536023054755,
          "macro avg": {
            "precision": 0.8301542207792207,
            "recall": 0.8860246495581804,
            "f1-score": 0.8528218858521326,
            "support": 1388.0
          },
          "weighted avg": {
            "precision": 0.9097827154833639,
            "recall": 0.8962536023054755,
            "f1-score": 0.9002857245226293,
            "support": 1388.0
          }
        },
        "roc_auc": 0.9566660895438165,
        "average_precision": 0.8271081453687152
      }
    }
  ]
}