• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1Quantization Backend Configuration
2----------------------------------
3
4FX Graph Mode Quantization allows the user to configure various
5quantization behaviors of an op in order to match the expectation
6of their backend.
7
8In the future, this document will contain a detailed spec of
9these configurations.
10
11
12Default values for native configurations
13~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
14
15Below is the output of the configuration for quantization of ops
16in x86 and qnnpack (PyTorch's default quantized backends).
17
18Results:
19
20.. literalinclude:: scripts/quantization_backend_configs/default_backend_config.txt
21