Contents Menu Expand Light mode Dark mode Auto light/dark mode
onnx-extended 0.2.3 documentation
Logo
onnx-extended 0.2.3 documentation

Contents

  • Tutorial
    • Build
      • Build with cython
      • Build with pybind11
      • Build with CUDA
      • Build with onnxruntime
    • CReferenceEvaluator
    • Cython Binding of onnxruntime
    • Custom Kernels for onnxruntime
    • Many Tools
      • External Data and Big Models
      • Onnx Manipulations
      • Quantization
      • Profiling onnxruntime
      • Debug Intermediate Results
      • Compare multiple versions of onnxruntime
      • Trees
    • Examples
      • Measuring CPU performance
      • Measuring CPU performance with a vector sum
      • Measuring CPU performance with a parallelized vector sum
      • Measuring CPU performance with a parallelized vector sum and AVX
      • Measuring CPU/GPU performance with a vector sum
    • Conv
      • Using C implementation of operator Conv
      • How float format has an impact on speed computation
      • Measuring Gemm performance with different input and output tests
      • Measuring performance about Gemm with onnxruntime
      • Profiles a simple onnx graph including a singleGemm
      • TreeEnsemble optimization
  • API
    • check
    • ext_test_case
    • helper
    • ortcy
    • ortops
      • ortops.tutorial
      • ortops.optim
    • reference
    • validation.cpu
    • validation.cuda
    • validation.bench_trees
    • tools
      • Shortcuts
      • tools.graph
      • tools.graph.onnx_graph_transformer
      • tools.onnx_nodes
      • Other tools
    • command lines
  • Technical Details
    • Install CUDA on WSL (2)
    • Useful commands on Linux
    • Gemm and storage order
    • 2023-09-05 - version GLIBCXX_3.4.30 not found
  • ONNX Benchmarks
  • Examples Gallery
    • Measuring CPU performance with a vector sum
    • Measuring CPU performance with a parallelized vector sum
    • Measuring CPU performance with a parallelized vector sum and AVX
    • Measuring CPU performance
    • Using C implementation of operator Conv
    • Measuring onnxruntime performance against a cython binding
    • Measuring CPU/GPU performance with a vector sum
    • Measuring Gemm performance with different input and output tests
    • How float format has an impact on speed computation
    • TreeEnsemble optimization
    • Profiles a simple onnx graph including a singleGemm
    • Measuring performance about Gemm with onnxruntime

More

  • Change Logs
  • LICENSE
Back to top

Tutorial#

This package is mostly used to validate custom implementations of a specific onnx operator or kernel. The first sections describe how to build the package and to add a new implementation depending the technology it relies on (CPU, openmp, CUDA, eigen, …). The last section is a sorted index of the examples.

Introduction#

  • Build
  • CReferenceEvaluator
  • Cython Binding of onnxruntime
  • Custom Kernels for onnxruntime
  • Many Tools
  • Examples
  • Conv
Next
Build
Previous
Home
Copyright © 2023, Xavier Dupré
Made with Sphinx and @pradyunsg's Furo
On this page
  • Tutorial
    • Introduction