CoolFace
Modelpublic

whr778/gliner2-multi-v1-cmnee

sourceHugging Faceotherupdated 2mo agoView on Hugging Face
0likes19downloads
Model Card

gliner2-multi-v1_cmnee

A GLiNER2 multi-task information-extraction model (entities, relations, events, and classification) fine-tuned from fastino/gliner2-multi-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-multi-v1`
  • —Library: gliner2
  • —Tasks: entity, relation, event, and classification extraction
  • —Experiment: gliner2-multi-v1_cmnee

Training data

1 dataset used for this run. 9,284 training records (val: 1,606, test: 2,727).

DatasetTask(s)TrainValTestLanguageLicenseSource
CMNEEEvent extraction (Chinese military)9,2841,6062,727zhsee sourcelink

Dataset notes

  • —CMNEE — Chinese military news event extraction; 8 event types, 11 argument roles.

Training procedure

SettingValue
Trained on2026-07-19
Duration1h 17m
Throughput17.9 samples/s
Epochs15
Batch size16 (× 2 grad-accum)
Encoder LR1e-05
Task-head LR0.0003
Weight decay0.01
Schedulercosine_restarts (warmup 0.05)
Precisionbf16
Max grad norm1.0
Best-checkpoint metricevaleventstrictmicrof1
Seed42

Evaluation

Decision threshold: 0.5 (config default).

Blind test (held-out test splits)

Micro precision / recall / F1, strict → relaxed.

CategoryPrecisionRecallF1Support
event_type1.000 → 1.0000.972 → 0.9720.986 → 0.9863819
event_trigger0.921 → 0.9240.832 → 0.8350.874 → 0.8774848
event_argument0.268 → 0.8380.188 → 0.6140.221 → 0.70918432
event0.533 → 0.8850.414 → 0.7090.466 → 0.78727099

Best checkpoint (validation)

Micro precision / recall / F1, strict → relaxed.

CategoryPrecisionRecallF1Support
event_type1.000 → 1.0000.981 → 0.9810.990 → 0.9902248
event_trigger0.894 → 0.9140.837 → 0.8560.864 → 0.8842643
event_argument0.308 → 0.8160.227 → 0.6230.261 → 0.7069366
event0.565 → 0.8720.459 → 0.7260.506 → 0.79214257

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-multi-v1 — see model card

Unverified — verify the upstream terms before redistribution

  • —CMNEE (see source)
  • —gliner2-multi-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).