4 ms·
Many businesses/services can't saturate the hardware you describe. It's just too much compute power. With CPUs you can scale down to fit your actual needs: all
by 37ef_ced3 5y ago
Many businesses/services can't saturate the hardware you describe. It's just too much compute power. With CPUs you can scale down to fit your actual needs: all the way down to a single AVX-512 core doing maybe 24 inferences per second (costing a few dollars PER MONTH).
Also, your cost/inference results will change if you use a fast CPU inference engine, instead of something slow like PyTorch (which you appear to be using).
- carbocation 5y agoThanks - this is something I wasn’t familiar with. Do you have any pointers for CPU inference engines that you’ve had good experience with or that I can look into further?
- markurtz 5y agoDisclosure: I work for Neural Magic. Hi carbocation, we'd love to see what you think of the performance using the DeepSparse engine for CPU inference: https://github.com/neuralmagic/deepsparse https://github.com/neuralmagic/deepsparse Take a look through our getting started pages that walk through performance benchmarking, training, and deployment for our featured models: https://sparsezoo.neuralmagic.com/getting-started https://sparsezoo.neuralmagic.com/getting-started