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]
ontonotes5
quynong/ontonotes5
18
[ "CARDINAL", "DATE", "PERSON", "NORP", "GPE", "LAW", "ORG", "PERCENT", "ORDINAL", "MONEY", "WORK_OF_ART", "FAC", "TIME", "LOC", "QUANTITY", "PRODUCT", "EVENT", "LANGUAGE" ]
test
[ "corenlp", "deberta-v3-base", "distilbert-base-cased" ]
[ 42, 43, 44 ]
5
16
null
null
{ "deberta-v3-base-seed42": { "status": "ok", "name": "deberta-v3-base", "seed": 42, "repo": "AITeamUIT/baseline-deberta-v3-base-ontonotes5-seed42", "summary": { "micro_precision": 0.9049, "micro_recall": 0.943, "micro_f1": 0.9235, "macro_precision": 0.8349, "macro_re...
[ { "baseline": "deberta-v3-base", "n": 3, "micro_f1_mean": 92.32, "micro_f1_std": 0.35, "macro_f1_mean": 85.16, "macro_f1_std": 0.95, "micro_p_mean": 90.79, "micro_r_mean": 93.9, "micro_f1": "92.32 +/- 0.35", "macro_f1": "85.16 +/- 0.95" }, { "baseline": "distilbert-ba...
2026-08-24T21:04:45

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

Check out the documentation for more information.

ontonotes5 — NER baselines (Stanford CoreNLP CRF, DeBERTa-v3-base, DistilBERT-base)

  • dataset: quynong/ontonotes5 (18 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
deberta-v3-base 3 92.32 +/- 0.35 85.16 +/- 0.95 90.79 93.9
distilbert-base-cased 3 90.83 +/- 0.29 81.75 +/- 1.14 89.2 92.52

Checkpoint tung run

generated 2026-08-24T21:04:45

Downloads last month
34

Collection including AITeamUIT/eval-baselines-ontonotes5