AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Causes and Effects of Unanticipated Numerical Deviations in Neural Network Inference Frameworks

Alexander Schlögl, Nora Hofer, Rainer Böhme

NeurIPS 2023 · Architectures and efficiency › CNN design · approach: Theory of convolutional inductive bias · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
cast model from float32 to float16 (ablation)Cifar10-R18 · 75 platformsnumber of EQCs: 9 -> 3Improved
cast model from float32 to float64 (ablation)Cifar10-R18 · 75 platformsnumber of EQCs: 9 -> 4Improved
increase CPU core count from 2 to 4 (ablation)Intel Ice Lake · Cifar10-R18number of EQCs: 5 -> 4Improved
increase CPU core count from 8 to 16 (ablation)Intel platforms · Cifar10-R18number of EQCs: 4 -> 3Improved
increase number of core counts beyond 16 (ablation)Intel platforms · Cifar10-R18number of EQCsNo clear effect
increase number of multiplications per convolution layer (ablation)convolution layer · remaining precisionremaining precision: 23 -> 0Hurt
increase number of multiplications from 10^6-10^7 to 10^8 (ablation)convolution layer · number of EQCsnumber of EQCsMixed
change weight/input distribution (original, Glorot, Gaussian; captured, permuted, uniform inputs) (ablation)conv2d_11 · dual-core GCP instancesnumber of EQCsNo clear effect
enable XLA compute-graph optimization (not activated)TensorFlow · CPU/GPU inferencenumber of EQCsNo clear effect
different device placementTensorFlow · CPU/GPU inferencenumber of EQCsNo clear effect
AMD Rome/Milan or Sandy/Ivy Bridge at larger core countsCPU inference · Cifar10-R18number of EQCsNo clear effect
different GPU model (GTX 1650 vs RTX 2070)Cifar10-R18 · GPU inferencenumber of EQCs: 1 -> 2Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy