NIST
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
NIST16
92.7F1 Score
93
NIST SD27 (test)
71F1 Score
77
NIST 2003 (MT-03)
50.3BLEU
52
NIST MT-05 2005
50.72BLEU
42
NIST
73.7F1 Score
34
NIST 16
0.872AUC
31
NIST
44IoU
30
NIST CTS CallHome 2000
11WER
27
NIST MT 06 2006 (test)
49.98BLEU
27
NIST MT 04 2004 (test)
0.506BLEU
27
NIST 2020 (Scaffold)
46Top-1 Accuracy
16
NIST 16
0.292P-F1
15
NIST16
0.76AUC
15
NIST MT-06
38.04BLEU
15
NIST MT 2004
40.6BLEU
15
NIST 16
54.9F1 Score
14
NIST16
0.47F1 (best)
14
NIST 2020 (Random split)
53.3Top-1 Accuracy
12
NIST 16
85.1AUC
12
NIST16 (test)
97.4F1 Score
12
NIST 16
0.5F1 Score
11
NIST16
72.71F1 Score
11
NIST Chinese-English MT08 (test)
40.45BLEU
11
NIST (test)
84Similarity (Sim_k)
10
NIST 16
0.447IoU
10