Authors apply a consistent methodology to benchmark physical performance metrics across neural network architectures and device technologies, identifying promising combinations.
The elements at the input, synapses, receive vectors of input signals, xi, and multiply them by vectors of weights, wi
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.ET 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Benchmarking Physical Performance of Neural Inference Circuits
Authors apply a consistent methodology to benchmark physical performance metrics across neural network architectures and device technologies, identifying promising combinations.