INDEX
Explanations
crucial metric, surprisingly effective, memory footprint
New Auto-Interp
Negative Logits
περισ
0.88
sterling
0.87
funzione
0.84
여행
0.83
funkcjon
0.82
czynności
0.82
ностран
0.82
petty
0.80
servicemen
0.80
ಭಯ
0.80
POSITIVE LOGITS
architectures
1.52
overfitting
1.43
computational
1.33
tractable
1.31
empirically
1.30
hyperparameters
1.29
GPUs
1.29
AutoML
1.27
convolutional
1.27
heuristics
1.26
Activations Density 1.219%