Skip to content

Pytorch w8a8 SalsaNext quantsim model
Compare
Choose a tag to compare
@quic-bharathr quic-bharathr released this 12 Feb 07:50
· 7 commits to develop since this release
torch_salsanext_models
59640d1

Optimized w8a8 checkpoint, encoding and FP32 checkpoint for Pytorch SalsaNext model.

For w8a8 optimization:

  • Batch norm folding followed by AdaRound in per channel mode have been applied on the original FP32 model.
  • Percentile was used in per channel mode for quantsim.
  • one operator activation output is enabled with 16-bit width.
Assets 5
Loading