Loaded 4386 rows from /root/workspace/tests/test.jsonl
Input keys: ['object', 'relation', 'sentence', 'sentence_id', 'subject', 'triple_id']
Unique occurrences: 7497
Candidate pairs: 5264
Unresolved rows before inference: 0
Candidate-count distribution: Counter({1: 3807, 2: 439, 3: 60, 4: 51, 5: 13, 6: 8, 8: 3, 9: 2, 15: 1, 7: 1, 18: 1})
Subject match modes: Counter({'exact': 4386})
Object match modes: Counter({'exact': 4386})
Using existence threshold from RE release config: 0.5
WojoodRelationPipeline(type_model_id='U4RASD/TypePredictor', re_model_id='U4RASD/DRU-RE-EntityPair-TwoHead-ARBERTv2', device='cuda', existence_threshold=0.5, candidate_positive_fraction=0.5, use_ontology_filter=True)
Encoded 500/7497 unique entity occurrences for TypePredictor
Encoded 1000/7497 unique entity occurrences for TypePredictor
Encoded 1500/7497 unique entity occurrences for TypePredictor
Encoded 2000/7497 unique entity occurrences for TypePredictor
Encoded 2500/7497 unique entity occurrences for TypePredictor
Encoded 3000/7497 unique entity occurrences for TypePredictor
Encoded 3500/7497 unique entity occurrences for TypePredictor
Encoded 4000/7497 unique entity occurrences for TypePredictor
Encoded 4500/7497 unique entity occurrences for TypePredictor
Encoded 5000/7497 unique entity occurrences for TypePredictor
Encoded 5500/7497 unique entity occurrences for TypePredictor
Encoded 6000/7497 unique entity occurrences for TypePredictor
Encoded 6500/7497 unique entity occurrences for TypePredictor
Encoded 7000/7497 unique entity occurrences for TypePredictor
You're using a PreTrainedTokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding.
TypePredictor: 1920/7497 occurrences
TypePredictor: 3840/7497 occurrences
TypePredictor: 5760/7497 occurrences
TypePredictor: 7497/7497 occurrences
Type prediction completed in 0.26 min
Encoded 1000/5264 candidate pairs for RE
Encoded 2000/5264 candidate pairs for RE
Encoded 3000/5264 candidate pairs for RE
Encoded 4000/5264 candidate pairs for RE
Encoded 5000/5264 candidate pairs for RE
You're using a BertTokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding.
Relation model: 1920/5264 candidates
Relation model: 3840/5264 candidates
Relation model: 5264/5264 candidates
Relation prediction completed in 0.09 min
max_joint label counts: [('Location.located_in', 1916), ('PartOf.geopolitical_division', 808), ('Location.lives_in', 406), ('no_relation', 314), ('Personal.has_occupation', 290), ('PartOf.subsidiary', 96), ('Affiliation.employee_of', 65), ('Administration.president_of', 65), ('Affiliation.member_of', 64), ('Business.has_partner_with', 63), ('Organization.has_alternate_name', 62), ('Personal.birth_place', 45), ('Business.has_conflict_with', 44), ('Administration.manager_of', 35), ('Location.nearby', 29), ('Organization.found_on', 11), ('Family.has_parent', 9), ('Organization.has_propoerty', 9), ('Location.headquartered_in', 9), ('Personal.death_date', 8), ('Affiliation.owner_of', 6), ('Productivity.builder_of', 4), ('Family.has_sibling', 4), ('Personal.birth_date', 3), ('GPE.has_currency', 3), ('Organization.has_revenue', 3), ('Productivity.founder_of', 2), ('GPE.official_language', 2), ('Affiliation.student_at', 2), ('Organization.employs', 2), ('Organization.branch_count', 2), ('Location.has_border_with', 2), ('Administration.leader_of', 1), ('Business.has_competitor', 1), ('GPE.has_area', 1)]
legacy_majority_sum label counts: [('Location.located_in', 1898), ('PartOf.geopolitical_division', 808), ('Location.lives_in', 406), ('no_relation', 357), ('Personal.has_occupation', 290), ('PartOf.subsidiary', 95), ('Affiliation.employee_of', 65), ('Affiliation.member_of', 64), ('Administration.president_of', 63), ('Organization.has_alternate_name', 60), ('Business.has_partner_with', 60), ('Personal.birth_place', 45), ('Business.has_conflict_with', 41), ('Administration.manager_of', 33), ('Location.nearby', 29), ('Organization.found_on', 11), ('Organization.has_propoerty', 9), ('Personal.death_date', 8), ('Family.has_parent', 7), ('Affiliation.owner_of', 6), ('Productivity.builder_of', 4), ('Personal.birth_date', 3), ('Family.has_sibling', 3), ('GPE.has_currency', 3), ('Location.headquartered_in', 3), ('Organization.has_revenue', 3), ('Productivity.founder_of', 2), ('Affiliation.student_at', 2), ('Organization.employs', 2), ('Organization.branch_count', 2), ('Location.has_border_with', 2), ('Administration.leader_of', 1), ('GPE.has_area', 1)]
top_existence label counts: [('Location.located_in', 1915), ('PartOf.geopolitical_division', 808), ('Location.lives_in', 407), ('no_relation', 314), ('Personal.has_occupation', 290), ('PartOf.subsidiary', 96), ('Affiliation.employee_of', 65), ('Administration.president_of', 65), ('Affiliation.member_of', 64), ('Business.has_partner_with', 63), ('Organization.has_alternate_name', 62), ('Personal.birth_place', 45), ('Business.has_conflict_with', 44), ('Administration.manager_of', 35), ('Location.nearby', 29), ('Organization.found_on', 11), ('Family.has_parent', 9), ('Organization.has_propoerty', 9), ('Location.headquartered_in', 9), ('Personal.death_date', 8), ('Affiliation.owner_of', 6), ('Productivity.builder_of', 4), ('Family.has_sibling', 4), ('Personal.birth_date', 3), ('GPE.has_currency', 3), ('Organization.has_revenue', 3), ('Productivity.founder_of', 2), ('GPE.official_language', 2), ('Affiliation.student_at', 2), ('Organization.employs', 2), ('Organization.branch_count', 2), ('Location.has_border_with', 2), ('Administration.leader_of', 1), ('Business.has_competitor', 1), ('GPE.has_area', 1)]
any_positive_max_q label counts: [('Location.located_in', 1916), ('PartOf.geopolitical_division', 808), ('Location.lives_in', 405), ('no_relation', 314), ('Personal.has_occupation', 290), ('PartOf.subsidiary', 96), ('Affiliation.employee_of', 65), ('Administration.president_of', 65), ('Affiliation.member_of', 64), ('Business.has_partner_with', 63), ('Organization.has_alternate_name', 62), ('Personal.birth_place', 45), ('Business.has_conflict_with', 45), ('Administration.manager_of', 35), ('Location.nearby', 29), ('Organization.found_on', 11), ('Family.has_parent', 9), ('Organization.has_propoerty', 9), ('Location.headquartered_in', 9), ('Personal.death_date', 8), ('Affiliation.owner_of', 6), ('Productivity.builder_of', 4), ('Family.has_sibling', 4), ('Personal.birth_date', 3), ('GPE.has_currency', 3), ('Organization.has_revenue', 3), ('Productivity.founder_of', 2), ('GPE.official_language', 2), ('Affiliation.student_at', 2), ('Organization.employs', 2), ('Organization.branch_count', 2), ('Location.has_border_with', 2), ('Administration.leader_of', 1), ('Business.has_competitor', 1), ('GPE.has_area', 1)]
soft_joint_pool label counts: [('Location.located_in', 1915), ('PartOf.geopolitical_division', 809), ('Location.lives_in', 405), ('no_relation', 314), ('Personal.has_occupation', 290), ('PartOf.subsidiary', 96), ('Affiliation.employee_of', 65), ('Administration.president_of', 65), ('Affiliation.member_of', 64), ('Business.has_partner_with', 63), ('Organization.has_alternate_name', 62), ('Personal.birth_place', 45), ('Business.has_conflict_with', 45), ('Administration.manager_of', 35), ('Location.nearby', 29), ('Organization.found_on', 11), ('Family.has_parent', 9), ('Organization.has_propoerty', 9), ('Location.headquartered_in', 9), ('Personal.death_date', 8), ('Affiliation.owner_of', 6), ('Productivity.builder_of', 4), ('Family.has_sibling', 4), ('Personal.birth_date', 3), ('GPE.has_currency', 3), ('Organization.has_revenue', 3), ('Productivity.founder_of', 2), ('GPE.official_language', 2), ('Affiliation.student_at', 2), ('Organization.employs', 2), ('Organization.branch_count', 2), ('Location.has_border_with', 2), ('Administration.leader_of', 1), ('Business.has_competitor', 1), ('GPE.has_area', 1)]

DONE
Primary aggregation: max_joint
Validated submission: /root/workspace/DRU-RE-EntityPair-TwoHead-ARBERTv2/test_codabench/submission.zip
RunPod convenience copy: /root/workspace/DRU-RE-EntityPair-TwoHead-ARBERTv2/test_codabench/submission.zip
Debug predictions: /root/workspace/DRU-RE-EntityPair-TwoHead-ARBERTv2/test_codabench/predictions_debug.jsonl
Manifest: /root/workspace/DRU-RE-EntityPair-TwoHead-ARBERTv2/test_codabench/run_manifest.json