{
  "created_at": "2026-05-18T17:47:01.228362+00:00",
  "config": {
    "hf_dataset": "faodl/amis-agri-trade-pri-sec",
    "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-trade-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-trade-classifier",
    "hub_private_repo": true
  },
  "dataset_summary": {
    "train": {
      "rows": 4799,
      "labels": {
        "0": 2363,
        "1": 2436
      },
      "unique_groups": 2483,
      "text_length_mean": 695.5259429047718,
      "text_length_median": 794.0
    },
    "validation": {
      "rows": 1009,
      "labels": {
        "0": 462,
        "1": 547
      },
      "unique_groups": 532,
      "text_length_mean": 698.1397423191279,
      "text_length_median": 794.0
    },
    "test": {
      "rows": 1017,
      "labels": {
        "0": 529,
        "1": 488
      },
      "unique_groups": 533,
      "text_length_mean": 694.6155358898721,
      "text_length_median": 794.0
    }
  },
  "results": [
    {
      "model_type": "logistic_tfidf",
      "model_name": "logistic",
      "artifact_dir": "/content/agri-trade-classifier/baselines/logistic",
      "artifact_file": "/content/agri-trade-classifier/baselines/logistic/logistic_tfidf.joblib",
      "validation_best_threshold": {
        "threshold": 0.39574905324813764,
        "f1": 0.8111658456486042,
        "precision": 0.736214605067064,
        "recall": 0.903107861060329
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.7384464110127827,
        "precision": 0.7361702127659574,
        "recall": 0.7090163934426229,
        "f1": 0.7223382045929019,
        "confusion_matrix": [
          [
            405,
            124
          ],
          [
            142,
            346
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.7404021937842779,
            "recall": 0.7655954631379962,
            "f1-score": 0.7527881040892194,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.7361702127659574,
            "recall": 0.7090163934426229,
            "f1-score": 0.7223382045929019,
            "support": 488.0
          },
          "accuracy": 0.7384464110127827,
          "macro avg": {
            "precision": 0.7382862032751176,
            "recall": 0.7373059282903096,
            "f1-score": 0.7375631543410606,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.7383715086938744,
            "recall": 0.7384464110127827,
            "f1-score": 0.7381769428756472,
            "support": 1017.0
          }
        },
        "roc_auc": 0.8378765223589203,
        "average_precision": 0.8147823236546703
      },
      "test_optimal_threshold": {
        "threshold": 0.39574905324813764,
        "accuracy": 0.744346116027532,
        "precision": 0.6743119266055045,
        "recall": 0.9036885245901639,
        "f1": 0.7723292469352014,
        "confusion_matrix": [
          [
            316,
            213
          ],
          [
            47,
            441
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.8705234159779615,
            "recall": 0.5973534971644613,
            "f1-score": 0.7085201793721974,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.6743119266055045,
            "recall": 0.9036885245901639,
            "f1-score": 0.7723292469352014,
            "support": 488.0
          },
          "accuracy": 0.744346116027532,
          "macro avg": {
            "precision": 0.772417671291733,
            "recall": 0.7505210108773126,
            "f1-score": 0.7404247131536994,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.7763727701433902,
            "recall": 0.744346116027532,
            "f1-score": 0.7391384930110824,
            "support": 1017.0
          }
        },
        "roc_auc": 0.8378765223589203,
        "average_precision": 0.8147823236546703
      }
    },
    {
      "model_type": "xgboost_tfidf",
      "model_name": "xgboost",
      "artifact_dir": "/content/agri-trade-classifier/baselines/xgboost",
      "artifact_file": "/content/agri-trade-classifier/baselines/xgboost/xgboost_tfidf.joblib",
      "validation_best_threshold": {
        "threshold": 0.30486294627189636,
        "f1": 0.8129139072847682,
        "precision": 0.7428139183055976,
        "recall": 0.8976234003656307
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.7620452310717798,
        "precision": 0.786046511627907,
        "recall": 0.6926229508196722,
        "f1": 0.7363834422657952,
        "confusion_matrix": [
          [
            437,
            92
          ],
          [
            150,
            338
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.7444633730834753,
            "recall": 0.8260869565217391,
            "f1-score": 0.7831541218637993,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.786046511627907,
            "recall": 0.6926229508196722,
            "f1-score": 0.7363834422657952,
            "support": 488.0
          },
          "accuracy": 0.7620452310717798,
          "macro avg": {
            "precision": 0.7652549423556911,
            "recall": 0.7593549536707056,
            "f1-score": 0.7597687820647973,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.7644167374981092,
            "recall": 0.7620452310717798,
            "f1-score": 0.7607115538757698,
            "support": 1017.0
          }
        },
        "roc_auc": 0.8467588862375656,
        "average_precision": 0.816020877643838
      },
      "test_optimal_threshold": {
        "threshold": 0.30486294627189636,
        "accuracy": 0.7522123893805309,
        "precision": 0.6849529780564263,
        "recall": 0.8954918032786885,
        "f1": 0.7761989342806395,
        "confusion_matrix": [
          [
            328,
            201
          ],
          [
            51,
            437
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.8654353562005277,
            "recall": 0.6200378071833649,
            "f1-score": 0.7224669603524229,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.6849529780564263,
            "recall": 0.8954918032786885,
            "f1-score": 0.7761989342806395,
            "support": 488.0
          },
          "accuracy": 0.7522123893805309,
          "macro avg": {
            "precision": 0.775194167128477,
            "recall": 0.7577648052310266,
            "f1-score": 0.7493329473165311,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.7788322091657967,
            "recall": 0.7522123893805309,
            "f1-score": 0.7482498544300726,
            "support": 1017.0
          }
        },
        "roc_auc": 0.8467588862375656,
        "average_precision": 0.816020877643838
      }
    },
    {
      "model_type": "embedding-logistic_sentence_embeddings",
      "model_name": "logistic",
      "embedding_model_name": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
      "artifact_dir": "/content/agri-trade-classifier/baselines/embedding-logistic",
      "artifact_file": "/content/agri-trade-classifier/baselines/embedding-logistic/embedding-logistic.joblib",
      "validation_best_threshold": {
        "threshold": 0.31512747482812214,
        "f1": 0.8587943848059455,
        "precision": 0.7831325301204819,
        "recall": 0.9506398537477148
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.7895771878072763,
        "precision": 0.75,
        "recall": 0.8422131147540983,
        "f1": 0.7934362934362934,
        "confusion_matrix": [
          [
            392,
            137
          ],
          [
            77,
            411
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.835820895522388,
            "recall": 0.7410207939508506,
            "f1-score": 0.7855711422845691,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.75,
            "recall": 0.8422131147540983,
            "f1-score": 0.7934362934362934,
            "support": 488.0
          },
          "accuracy": 0.7895771878072763,
          "macro avg": {
            "precision": 0.792910447761194,
            "recall": 0.7916169543524745,
            "f1-score": 0.7895037178604313,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.7946403674841134,
            "recall": 0.7895771878072763,
            "f1-score": 0.7893451774488184,
            "support": 1017.0
          }
        },
        "roc_auc": 0.8809306145216771,
        "average_precision": 0.8633117674086846
      },
      "test_optimal_threshold": {
        "threshold": 0.31512747482812214,
        "accuracy": 0.7708947885939036,
        "precision": 0.6976744186046512,
        "recall": 0.9221311475409836,
        "f1": 0.794351279788173,
        "confusion_matrix": [
          [
            334,
            195
          ],
          [
            38,
            450
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.8978494623655914,
            "recall": 0.6313799621928167,
            "f1-score": 0.7413984461709212,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.6976744186046512,
            "recall": 0.9221311475409836,
            "f1-score": 0.794351279788173,
            "support": 488.0
          },
          "accuracy": 0.7708947885939036,
          "macro avg": {
            "precision": 0.7977619404851213,
            "recall": 0.7767555548669001,
            "f1-score": 0.7678748629795471,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.8017969339925934,
            "recall": 0.7708947885939036,
            "f1-score": 0.766807475477921,
            "support": 1017.0
          }
        },
        "roc_auc": 0.8809306145216771,
        "average_precision": 0.8633117674086846
      }
    },
    {
      "model_type": "embedding-svm_sentence_embeddings",
      "model_name": "svm",
      "embedding_model_name": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
      "artifact_dir": "/content/agri-trade-classifier/baselines/embedding-svm",
      "artifact_file": "/content/agri-trade-classifier/baselines/embedding-svm/embedding-svm.joblib",
      "validation_best_threshold": {
        "threshold": 0.45313322034755105,
        "f1": 0.8613264427217915,
        "precision": 0.8143322475570033,
        "recall": 0.9140767824497258
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.7876106194690266,
        "precision": 0.7419928825622776,
        "recall": 0.8545081967213115,
        "f1": 0.7942857142857143,
        "confusion_matrix": [
          [
            384,
            145
          ],
          [
            71,
            417
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.843956043956044,
            "recall": 0.725897920604915,
            "f1-score": 0.7804878048780488,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.7419928825622776,
            "recall": 0.8545081967213115,
            "f1-score": 0.7942857142857143,
            "support": 488.0
          },
          "accuracy": 0.7876106194690266,
          "macro avg": {
            "precision": 0.7929744632591608,
            "recall": 0.7902030586631132,
            "f1-score": 0.7873867595818815,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.7950297678890252,
            "recall": 0.7876106194690266,
            "f1-score": 0.7871086306311862,
            "support": 1017.0
          }
        },
        "roc_auc": 0.883096005454151,
        "average_precision": 0.8650284607518275
      },
      "test_optimal_threshold": {
        "threshold": 0.45313322034755105,
        "accuracy": 0.7964601769911505,
        "precision": 0.7353433835845896,
        "recall": 0.8995901639344263,
        "f1": 0.8092165898617512,
        "confusion_matrix": [
          [
            371,
            158
          ],
          [
            49,
            439
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.8833333333333333,
            "recall": 0.7013232514177694,
            "f1-score": 0.7818756585879874,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.7353433835845896,
            "recall": 0.8995901639344263,
            "f1-score": 0.8092165898617512,
            "support": 488.0
          },
          "accuracy": 0.7964601769911505,
          "macro avg": {
            "precision": 0.8093383584589615,
            "recall": 0.8004567076760978,
            "f1-score": 0.7955461242248693,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.8123214400419008,
            "recall": 0.7964601769911505,
            "f1-score": 0.7949950041746116,
            "support": 1017.0
          }
        },
        "roc_auc": 0.883096005454151,
        "average_precision": 0.8650284607518275
      }
    },
    {
      "model_type": "embedding-lightgbm_sentence_embeddings",
      "model_name": "lightgbm",
      "embedding_model_name": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
      "artifact_dir": "/content/agri-trade-classifier/baselines/embedding-lightgbm",
      "artifact_file": "/content/agri-trade-classifier/baselines/embedding-lightgbm/embedding-lightgbm.joblib",
      "validation_best_threshold": {
        "threshold": 0.14811074908471936,
        "f1": 0.8659450457951707,
        "precision": 0.7951070336391437,
        "recall": 0.9506398537477148
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.7817109144542773,
        "precision": 0.7435897435897436,
        "recall": 0.8319672131147541,
        "f1": 0.7852998065764023,
        "confusion_matrix": [
          [
            389,
            140
          ],
          [
            82,
            406
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.8259023354564756,
            "recall": 0.7353497164461248,
            "f1-score": 0.778,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.7435897435897436,
            "recall": 0.8319672131147541,
            "f1-score": 0.7852998065764023,
            "support": 488.0
          },
          "accuracy": 0.7817109144542773,
          "macro avg": {
            "precision": 0.7847460395231096,
            "recall": 0.7836584647804394,
            "f1-score": 0.7816499032882012,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.7864052412274046,
            "recall": 0.7817109144542773,
            "f1-score": 0.7815027587111941,
            "support": 1017.0
          }
        },
        "roc_auc": 0.8798769717065914,
        "average_precision": 0.867377313554101
      },
      "test_optimal_threshold": {
        "threshold": 0.14811074908471936,
        "accuracy": 0.7590953785644051,
        "precision": 0.684931506849315,
        "recall": 0.9221311475409836,
        "f1": 0.7860262008733624,
        "confusion_matrix": [
          [
            322,
            207
          ],
          [
            38,
            450
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.8944444444444445,
            "recall": 0.6086956521739131,
            "f1-score": 0.7244094488188977,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.684931506849315,
            "recall": 0.9221311475409836,
            "f1-score": 0.7860262008733624,
            "support": 488.0
          },
          "accuracy": 0.7590953785644051,
          "macro avg": {
            "precision": 0.7896879756468798,
            "recall": 0.7654133998574484,
            "f1-score": 0.7552178248461301,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.7939111961195446,
            "recall": 0.7590953785644051,
            "f1-score": 0.753975795920745,
            "support": 1017.0
          }
        },
        "roc_auc": 0.8798769717065914,
        "average_precision": 0.867377313554101
      }
    },
    {
      "model_type": "transformer",
      "model_name": "FacebookAI/xlm-roberta-base",
      "artifact_dir": "/content/agri-trade-classifier/transformer",
      "validation_best_threshold": {
        "threshold": 0.383081316947937,
        "f1": 0.8737373737373737,
        "precision": 0.8096723868954758,
        "recall": 0.9488117001828154
      },
      "test_default_0_5": {
        "threshold": 0.5,
        "accuracy": 0.8367748279252704,
        "precision": 0.7864768683274022,
        "recall": 0.9057377049180327,
        "f1": 0.8419047619047619,
        "confusion_matrix": [
          [
            409,
            120
          ],
          [
            46,
            442
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.8989010989010989,
            "recall": 0.7731568998109641,
            "f1-score": 0.8313008130081301,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.7864768683274022,
            "recall": 0.9057377049180327,
            "f1-score": 0.8419047619047619,
            "support": 488.0
          },
          "accuracy": 0.8367748279252704,
          "macro avg": {
            "precision": 0.8426889836142506,
            "recall": 0.8394473023644984,
            "f1-score": 0.836602787456446,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.8449551554203084,
            "recall": 0.8367748279252704,
            "f1-score": 0.8363890402073005,
            "support": 1017.0
          }
        },
        "roc_auc": 0.9186951873314947,
        "average_precision": 0.9127183314490565
      },
      "test_optimal_threshold": {
        "threshold": 0.383081316947937,
        "accuracy": 0.8367748279252704,
        "precision": 0.7710437710437711,
        "recall": 0.9385245901639344,
        "f1": 0.8465804066543438,
        "confusion_matrix": [
          [
            393,
            136
          ],
          [
            30,
            458
          ]
        ],
        "classification_report": {
          "NOT_RELEVANT": {
            "precision": 0.9290780141843972,
            "recall": 0.7429111531190926,
            "f1-score": 0.8256302521008403,
            "support": 529.0
          },
          "RELEVANT": {
            "precision": 0.7710437710437711,
            "recall": 0.9385245901639344,
            "f1-score": 0.8465804066543438,
            "support": 488.0
          },
          "accuracy": 0.8367748279252704,
          "macro avg": {
            "precision": 0.8500608926140841,
            "recall": 0.8407178716415136,
            "f1-score": 0.836105329377592,
            "support": 1017.0
          },
          "weighted avg": {
            "precision": 0.8532464402880102,
            "recall": 0.8367748279252704,
            "f1-score": 0.8356830302936719,
            "support": 1017.0
          }
        },
        "roc_auc": 0.9186951873314947,
        "average_precision": 0.9127183314490565
      }
    }
  ]
}