Shakespeare
Benchmarks
TaskDatasetSOTA resultTrendResults
Shakespeare (val)
3.721Perplexity
73
Shakespeare modern
55.63Accuracy
48
Shakespeare's text
0.16Cross-Entropy
42
SHAKESPEARE
53Diversity Score
40
Shakespeare 1M params (val)
4.79NLL
35
Shakespeare
50.611Accuracy
33
Shakespeare (test)
47.55Accuracy
31
Shakespeare (val)
10.08Perplexity
27
Shakespeare
55.9Accuracy (Mean)
25
Shakespeare Low Resource
10.46PPL
15
Shakespeare High Resource
5.01PPL
15
Shakespeare (train)
59.8Accuracy
12
Shakespeare (test)
50.2Accuracy
12
Shakespeare realistic non-IID LEAF benchmark (test)
54.52Top-1 Acc
11
Shakespeare
0.17Epsilon LB
9
Tiny-Shakespeare (o.o.d)
569.3Perplexity
9
ShakeSpeare LEAF (non-i.i.d.)
46.36Accuracy (500R)
9
Shakespeare (test)
89.77Percent Clients Acc_p Greater Than Acc_g
8
Shakespeare (test)
89.77Accp > Accg %
8
Shakespeare (non-i.i.d.)
45.24Test Error
8
Shakespeare i.i.d.
44.52Test Error
8
Shakespeare (non-i.i.d.)
45.24Test Error
8
Shakespeare (i.i.d.)
44.52Test Error
8
Shakespeare (n=30, m=3) (test)
51.63Accuracy
7
Shakespeare n=15, m=7 (test)
53.67Test Accuracy
7