{ "status": "completed", "completed_at_utc": "2026-07-12T18:56:34+00:00", "dataset_repo_id": "U4RASD/TypePrediction", "model_repo_id": "U4RASD/TypePredictor", "architecture": { "architecture": "NeoAraBERTTypePredictor", "task": "mention-level 21-class Arabic entity typing", "base_model": "U4RASD/NeoAraBERT", "input_markers": [ "[ENT]", "[/ENT]" ], "representation": "last hidden state at token position 0 (CLS)", "dropout": 0.1, "classifier": [ 768, 21 ], "loss": "ordinary unweighted multiclass cross-entropy", "num_labels": 21, "labels": [ "GPE", "ORG", "DATE", "PERS", "NORP", "ORDINAL", "OCC", "EVENT", "CARDINAL", "LOC", "WEBSITE", "FAC", "LAW", "TIME", "MONEY", "CURR", "LANGUAGE", "PERCENT", "PRODUCT", "QUANTITY", "UNIT" ], "tokenizer_size": 65002, "hidden_size": 768, "max_length": 512, "context_candidates": [ null, 500, 300, 150, 80, 30, 0 ] }, "dataset_counts": { "train": 100796, "validation": 12600, "test": 12600 }, "validation_results": { "overall": { "subset": "overall", "rows": 12600, "types_present": [ "GPE", "ORG", "DATE", "PERS", "NORP", "ORDINAL", "OCC", "EVENT", "CARDINAL", "LOC", "WEBSITE", "FAC", "LAW", "TIME", "MONEY", "CURR", "LANGUAGE", "PERCENT", "PRODUCT", "QUANTITY", "UNIT" ], "num_types_present": 21, "accuracy": 0.9778571428571429, "micro_precision": 0.9778571428571429, "micro_recall": 0.9778571428571429, "micro_f1": 0.9778571428571429, "macro_precision_present_types": 0.9725024753902682, "macro_recall_present_types": 0.9475659044801065, "macro_f1_present_types": 0.9587281193044483, "macro_precision_all_21": 0.9725024753902682, "macro_recall_all_21": 0.9475659044801065, "macro_f1_all_21": 0.9587281193044483, "weighted_precision": 0.977879248064854, "weighted_recall": 0.9778571428571429, "weighted_f1": 0.9777906938936711, "split": "validation", "checkpoint": "best", "model_repo_id": "U4RASD/TypePredictor", "generated_at_utc": "2026-07-12T18:55:47+00:00" }, "unseen_sentence": { "subset": "unseen_sentence", "rows": 939, "types_present": [ "GPE", "ORG", "DATE", "PERS", "NORP", "ORDINAL", "OCC", "EVENT", "CARDINAL", "LOC", "WEBSITE", "FAC", "LAW", "TIME", "MONEY", "CURR", "LANGUAGE", "PERCENT", "PRODUCT", "UNIT" ], "num_types_present": 20, "accuracy": 0.9456869009584664, "micro_precision": 0.9456869009584664, "micro_recall": 0.9456869009584664, "micro_f1": 0.9456869009584664, "macro_precision_present_types": 0.9465091355300324, "macro_recall_present_types": 0.8850947519406029, "macro_f1_present_types": 0.9088413774156396, "macro_precision_all_21": 0.9014372719333641, "macro_recall_all_21": 0.8429473828005742, "macro_f1_all_21": 0.8655632165863235, "weighted_precision": 0.9458822212425712, "weighted_recall": 0.9456869009584664, "weighted_f1": 0.945010776949412, "split": "validation", "checkpoint": "best", "model_repo_id": "U4RASD/TypePredictor", "generated_at_utc": "2026-07-12T18:55:47+00:00" }, "seen_sentence_new_entity": { "subset": "seen_sentence_new_entity", "rows": 11661, "types_present": [ "GPE", "ORG", "DATE", "PERS", "NORP", "ORDINAL", "OCC", "EVENT", "CARDINAL", "LOC", "WEBSITE", "FAC", "LAW", "TIME", "MONEY", "CURR", "LANGUAGE", "PERCENT", "PRODUCT", "QUANTITY", "UNIT" ], "num_types_present": 21, "accuracy": 0.9804476459994854, "micro_precision": 0.9804476459994854, "micro_recall": 0.9804476459994854, "micro_f1": 0.9804476459994854, "macro_precision_present_types": 0.9751134735007793, "macro_recall_present_types": 0.9537701979295099, "macro_f1_present_types": 0.963262123278673, "macro_precision_all_21": 0.9751134735007793, "macro_recall_all_21": 0.9537701979295099, "macro_f1_all_21": 0.963262123278673, "weighted_precision": 0.980497781530163, "weighted_recall": 0.9804476459994854, "weighted_f1": 0.9804033222521132, "split": "validation", "checkpoint": "best", "model_repo_id": "U4RASD/TypePredictor", "generated_at_utc": "2026-07-12T18:55:47+00:00" } }, "test_results": { "overall": { "subset": "overall", "rows": 12600, "types_present": [ "GPE", "ORG", "DATE", "PERS", "NORP", "ORDINAL", "OCC", "EVENT", "CARDINAL", "LOC", "WEBSITE", "FAC", "LAW", "TIME", "MONEY", "CURR", "LANGUAGE", "PERCENT", "PRODUCT", "QUANTITY", "UNIT" ], "num_types_present": 21, "accuracy": 0.9793650793650793, "micro_precision": 0.9793650793650793, "micro_recall": 0.9793650793650793, "micro_f1": 0.9793650793650793, "macro_precision_present_types": 0.9672581856551346, "macro_recall_present_types": 0.9534679938651157, "macro_f1_present_types": 0.9599200012897606, "macro_precision_all_21": 0.9672581856551346, "macro_recall_all_21": 0.9534679938651157, "macro_f1_all_21": 0.9599200012897606, "weighted_precision": 0.9793254669511197, "weighted_recall": 0.9793650793650793, "weighted_f1": 0.9792974524068463, "split": "test", "checkpoint": "best", "model_repo_id": "U4RASD/TypePredictor", "generated_at_utc": "2026-07-12T18:56:34+00:00" }, "unseen_sentence": { "subset": "unseen_sentence", "rows": 994, "types_present": [ "GPE", "ORG", "DATE", "PERS", "NORP", "ORDINAL", "OCC", "EVENT", "CARDINAL", "LOC", "WEBSITE", "FAC", "LAW", "TIME", "MONEY", "CURR", "LANGUAGE", "PERCENT", "PRODUCT", "QUANTITY" ], "num_types_present": 20, "accuracy": 0.9627766599597586, "micro_precision": 0.9627766599597586, "micro_recall": 0.9627766599597586, "micro_f1": 0.9627766599597586, "macro_precision_present_types": 0.9052965403310622, "macro_recall_present_types": 0.886274373082116, "macro_f1_present_types": 0.8931735856196752, "macro_precision_all_21": 0.8621871812676783, "macro_recall_all_21": 0.8440708315067772, "macro_f1_all_21": 0.8506415101139764, "weighted_precision": 0.9615439764829642, "weighted_recall": 0.9627766599597586, "weighted_f1": 0.9617803672434098, "split": "test", "checkpoint": "best", "model_repo_id": "U4RASD/TypePredictor", "generated_at_utc": "2026-07-12T18:56:34+00:00" }, "seen_sentence_new_entity": { "subset": "seen_sentence_new_entity", "rows": 11606, "types_present": [ "GPE", "ORG", "DATE", "PERS", "NORP", "ORDINAL", "OCC", "EVENT", "CARDINAL", "LOC", "WEBSITE", "FAC", "LAW", "TIME", "MONEY", "CURR", "LANGUAGE", "PERCENT", "PRODUCT", "QUANTITY", "UNIT" ], "num_types_present": 21, "accuracy": 0.9807858004480441, "micro_precision": 0.9807858004480441, "micro_recall": 0.9807858004480441, "micro_f1": 0.9807858004480441, "macro_precision_present_types": 0.970282742754196, "macro_recall_present_types": 0.9589517514121707, "macro_f1_present_types": 0.9641131075953279, "macro_precision_all_21": 0.970282742754196, "macro_recall_all_21": 0.9589517514121707, "macro_f1_all_21": 0.9641131075953279, "weighted_precision": 0.9807493542827415, "weighted_recall": 0.9807858004480441, "weighted_f1": 0.9807123412635281, "split": "test", "checkpoint": "best", "model_repo_id": "U4RASD/TypePredictor", "generated_at_utc": "2026-07-12T18:56:34+00:00" } }, "checkpoint_summary": { "best_checkpoint_source": "outputs/TypePredictor/checkpoints/checkpoint-25000", "best_checkpoint_preserved": "outputs/TypePredictor/best_checkpoint", "best_step": 25000, "best_epoch": 3.9682923925552602, "best_validation_macro_f1": 0.9587281193044483, "latest_checkpoint_source": "outputs/TypePredictor/checkpoints/checkpoint-25196", "latest_checkpoint_preserved": "outputs/TypePredictor/latest_checkpoint", "latest_step": 25196, "latest_epoch": 3.9994047382832654, "best_and_latest_same_source": false, "root_released_model": "best checkpoint" }, "training_wall_seconds": 12262.467122793198, "train_metrics": { "train_runtime": 12261.7153, "train_samples_per_second": 32.882, "train_steps_per_second": 2.055, "total_flos": 0.0, "train_loss": 0.19069333035238045, "epoch": 3.9994047382832654 }, "trainer_log_history_entries": 554 }