SQS: Bayesian DNN Compression through Sparse Quantized Sub-distributions
Open paper
Benchmark
Model
Rank
Results
image-classification-on-cifar-100
SQS (ResNet-18,4bit,50%nonzero,M=50)
#181
Percentage correct: 76.63
image-classification-on-cifar-10
SQS (ResNet-56,2bit,50%nonzero)
#205
Percentage correct: 93.53
Rank counts only results with a code link.