Dataset Viewer
Auto-converted to Parquet Duplicate
dataset
string
dataset_repo
string
n_labels
int64
labels
list
eval_split
string
baselines
list
seeds
list
epochs_hf
int64
batch_size_hf
int64
max_train_samples
null
max_eval_samples
null
runs
dict
comparison_table
list
generated_at
timestamp[s]
fin
quynong/fin
4
[ "PER", "ORG", "LOC", "MISC" ]
test
[ "corenlp", "deberta-v3-base", "distilbert-base-cased" ]
[ 42, 43, 44 ]
5
16
null
null
{ "corenlp": { "status": "ok", "name": "corenlp", "seed": null, "repo": "AITeamUIT/baseline-corenlp-fin", "summary": { "micro_precision": 0.9194, "micro_recall": 0.7238, "micro_f1": 0.8099, "macro_precision": 0.8343, "macro_recall": 0.4567, "macro_f1": 0.5466, ...
[ { "baseline": "corenlp", "n": 0, "micro_f1_mean": 80.99, "micro_f1_std": null, "macro_f1_mean": 54.66, "macro_f1_std": null, "micro_p_mean": 91.94, "micro_r_mean": 72.38, "micro_f1": "80.99", "macro_f1": "54.66" }, { "baseline": "deberta-v3-base", "n": 3, "mic...
2026-08-24T15:13:50

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

fin — NER baselines (Stanford CoreNLP CRF, DeBERTa-v3-base, DistilBERT-base)

  • dataset: quynong/fin (4 nhan), eval tren split test
  • HF baselines: 5 epochs, batch 16, seeds [42, 43, 44]
  • corenlp: 1 lan train (CRF deterministic, khong seed)

Ket qua (P/R/F1 %, HungarianEvaluator IoU>=0.5 hoac fuzzy)

baseline n micro F1 macro F1 micro P micro R
corenlp 0 80.99 54.66 91.94 72.38
deberta-v3-base 3 81.67 +/- 0.34 54.83 +/- 0.36 76.08 88.15
distilbert-base-cased 3 77.30 +/- 1.63 51.20 +/- 1.35 74.32 80.53

Checkpoint tung run

generated 2026-08-24T15:13:50

Downloads last month
8

Collection including AITeamUIT/eval-baselines-fin