Named Entity Recognition (NER) on SLUE

Metric: label-F1 (%) (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	label-F1 (%)▼	Extra Data	Paper	Date↕	Code
1	W2V2-L-LL60K (pipeline approach, uses LM)	82.2	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
2	W2V2-B-LS960 (pipeline approach, uses LM)	79.8	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
3	W2V2-L-LL60K (pipeline approach)	78.8	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
4	W2V2-B-LS960 (pipeline approach)	74.2	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
5	W2V2-L-LL60K (e2e approach, uses LM)	73.3	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
6	W2V2-B-LS960 (e2e approach, uses LM)	71.7	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
7	HuBERT-B-LS960 (e2e approach, uses LM)	70.3	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
8	W2V2-B-VP100K (e2e approach, uses LM)	69.8	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
9	W2V2-L-LL60K (e2e approach)	64.7	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
10	W2V2-B-LS960 (e2e approach)	64	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
11	HuBERT-B-LS960 (e2e approach)	62.9	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code
12	W2V2-B-VP100K (e2e approach)	60.8	No	SLUE: New Benchmark Tasks for Spoken Language Un...	2021-11-19	Code

#1W2V2-L-LL60K (pipeline approach, uses LM)SOTA
82.2
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#2W2V2-B-LS960 (pipeline approach, uses LM)
79.8
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#3W2V2-L-LL60K (pipeline approach)
78.8
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#4W2V2-B-LS960 (pipeline approach)
74.2
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#5W2V2-L-LL60K (e2e approach, uses LM)
73.3
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#6W2V2-B-LS960 (e2e approach, uses LM)
71.7
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#7HuBERT-B-LS960 (e2e approach, uses LM)
70.3
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#8W2V2-B-VP100K (e2e approach, uses LM)
69.8
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#9W2V2-L-LL60K (e2e approach)
64.7
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#10W2V2-B-LS960 (e2e approach)
64
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#11HuBERT-B-LS960 (e2e approach)
62.9
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code
#12W2V2-B-VP100K (e2e approach)
60.8
label-F1 (%)· 2021-11-19
SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech Code