OpenNN Performance Benchmark
Compare OpenNN against other frmeworks
Higher Capacity
Manage Bigger Data Sets
On the same computer, our library OpenNN loads data sets 1.8 times bigger than TensorFlow and PyTorch.
Faster Training
Boost your productivity
On average, OpenNN trains a model 2.508 times faster than PyTorch and 1.551 times faster than TensorFlow.
Lower Energy Consumption
Save power, save money
A native C++ runtime, with no interpreter overhead and multithreaded BLAS-backed math, uses less CPU time per epoch.
Better Generalization
Smarter model selection
The buit-in ModelSelection class searches architectures and hyperparameters automatically, producing models with lower test error.