whr778/gliner2-base-v1-redocred

gliner2-base-v1_redocred

A GLiNER2 multi-task information-extraction model (entities, relations, events, and classification) fine-tuned from fastino/gliner2-base-v1.

⚠️ License at a glance

• Effective license: Unverified — review required
• Commercial use: Unverified
• All dataset licenses verified: No

See License for the full determination and per-dataset terms.

Model details

• Base model: fastino/gliner2-base-v1
• Library: gliner2
• Tasks: entity, relation, event, and classification extraction
• Experiment: gliner2-base-v1_redocred

Training data

1 dataset used for this run. 3,053 training records (val: 500, test: 500).

Dataset	Task(s)	Train	Val	Test	Language	License	Source
Re-DocRED	NER + relations (doc-level)	3,053	500	500	en	see card	link

Dataset notes

• Re-DocRED — Revised DocRED with corrected gold annotations; canonical train/val/test splits, 96 relation types, 6 entity types.

Training procedure

Setting	Value
Trained on	2026-07-19
Duration	49m 46s
Throughput	12.6 samples/s
Epochs	10
Batch size	8 (× 2 grad-accum)
Encoder LR	1e-05
Task-head LR	0.0003
Weight decay	0.01
Scheduler	cosine_restarts (warmup 0.05)
Precision	bf16
Max grad norm	1.0
Best-checkpoint metric	eval_relation_strict_micro_f1
Seed	42

Evaluation

Decision threshold: 0.5 (config default).

Blind test (held-out test splits)

Micro precision / recall / F1, strict → relaxed.

Category	Precision	Recall	F1	Support
entity	0.838 → 0.889	0.845 → 0.897	0.842 → 0.893	10705
relation	0.603 → 0.735	0.168 → 0.205	0.263 → 0.321	17348

Best checkpoint (validation)

Micro precision / recall / F1, strict → relaxed.

Category	Precision	Recall	F1	Support
entity	0.831 → 0.885	0.833 → 0.888	0.832 → 0.887	12255
relation	0.632 → 0.761	0.174 → 0.210	0.273 → 0.329	18321

License

Effective license: Unverified — review required. This model is a derivative of its base model and every training dataset, so the most restrictive term across all of them governs the whole model.

• Commercial use: Unverified
• Share-alike obligation: No
• All licenses verified: No
• Base model: gliner2-base-v1 — see model card

Unverified — verify the upstream terms before redistribution

• Re-DocRED (see card)
• gliner2-base-v1 (see model card)

License strings are copied verbatim from each dataset's card/source and from tools/train/dataset_registry.yaml. "see card"/"see source"/"other" mean the upstream declares no clear license — treat as unverified. This summary is informational, not legal advice; confirm terms before redistribution or commercial use.

Citation

If you use this model, please cite GLiNER2 and the underlying datasets (linked in Training data).

---

Model card generated automatically at the end of training (2026-07-19).