Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

PTB

Benchmarks

Task NameDataset NameSOTA ResultTrend
Language ModelingPTB
Perplexity4.345
1,034
Language ModelingPTB (test)
Perplexity8.159
526
ImputationPTB
PRD4.16
162
Constituent parsingPTB (test)
F196.48
127
Language ModelingPTB (val)
Perplexity20.26
99
Unsupervised ParsingPTB (test)
F1 Score83.3
75
Open-ended text generationPTB
COH Score67.7
64
Phrase-structure parsingPTB (§23)
F1 Score95.8
56
Character-level PredictionPTB (test)
BPC (Test)1.045
42
Dependency ParsingPTB
LAS96.3
31
Grammar InductionPTB English (test)
F1 Score84.3
29
POS taggingPTB (test)
Accuracy97.78
24
Language ModelingPTB zero-shot
Perplexity82.05
23
Word-level predictionPTB word-level (test)
Perplexity72.8
19
ParsingPTB (test)
Sents/sec1,092
17
Language ModelingPTB English Mikolov preprocessed (val)
Perplexity44.9
13
Constituency ParsingPTB Section 23 (test)
Rerank F10.9513
13
ECG ClassificationPTB database
Accuracy99.43
13
Uncertainty QuantificationPTB
CU417
12
Membership Inference AttackPTB
AUC59.28
12
Constituency ParsingPTB (test)
Speed (Sents/s)1,127
12
Medical Time Series ClassificationPTB-XL 5-Classes (test)
Accuracy0.7353
11
Medical Time Series ClassificationPTB 2-Classes (test)
Accuracy0.8596
11
Text GenerationPTB (test)
Grammaticality0.447
10
ECG ClassificationPTB-XL
Accuracy81.4
8
Showing 25 of 52 rows