Code

Implement quantization on-the-fly#100

Code

Draft

saharNooby wants to merge 6 commits intomasterRWKV/rwkv.cpp:masterfrom quantization-on-the-flyRWKV/rwkv.cpp:quantization-on-the-flyCopy head branch name to clipboard

+332-58

Commits

Commits on Jun 13, 2023

• Implement on-the-fly quantization

  saharNooby

  committed
  65cdb51

Commits on Jun 14, 2023

• Resolve TODO items

  saharNooby

  committed
  dca26e9

• Fix error code

  saharNooby

  committed
  7a13fd2

• Reformat code

  saharNooby

  committed
  4d27fa8

• Consistently use FP16 and FP32 for rwkv.cpp data types

  saharNooby

  committed
  d3b6749

• Add test for on-the-fly quantization

  saharNooby

  committed
  c49d3d8