1. a851bba COMPMID-1995: Improves DepthwiseConvolution for qasymm8. by Georgios Pinitas · 5 years ago
  2. 8f43d74 COMPMID-2063: New Winograd implementation by Pablo Tello · 5 years ago
  3. a4bba9c COMPMID-1995: Fix 32-bit NEDepthwiseConvolution errors. by Georgios Pinitas · 5 years ago
  4. 47d39dc COMPMID-1975: Update depthwise convolution. by Georgios Pinitas · 5 years ago
  5. d3d97d2 COMPMID-1623: NEWinograd reduce the number of output tiles. by Pablo Tello · 6 years ago
  6. eb027e9 COMPMID-1600: Reduce number of tile specialisations. by Pablo Tello · 6 years ago
  7. a799ce0 COMPMID-1564: Add NEDepthwiseConvolution3x3 for QASYMM8 by Georgios Pinitas · 6 years ago
  8. 20c246a COMPMID-1532: Add DepthwiseConvolution3x3 FP16 on NEON by Georgios Pinitas · 6 years ago
  9. 000d33a COMPMID-1552: support kernels sizes 1x7, 7x1, 1x5, 5x1 in NEWinograd by Pablo Tello · 6 years ago
  10. 72686fa COMPMID-1550: Winograd integrate RSH changes. by Pablo Tello · 6 years ago
  11. bda6e4b COMPMID-1247:Integrate kernel size 1x3 & 3x1 support in NEWinogradLayer. by Pablo Tello · 6 years ago
  12. 79ffade COMPMID-1112: Enabled multithreading transforms in Winograd. by Pablo Tello · 6 years ago
  13. b3b1e0b COMPMID-1010: Remove RSH profiler header by Georgios Pinitas · 6 years ago
  14. 4d9379a COMPMID-1040: Added support for nullptr bias tensor in NEWinogradLayer by Andrew Mundy · 6 years ago
  15. be0ae93 COMPMID-1005: Update Depthwise Convolution form RSH by Georgios Pinitas · 6 years ago
  16. a8a28f6 COMPMID-976: Fix warnings for ndk-r16b by Anthony Barbier · 6 years ago
  17. d267b05 COMPMID-784: Fixed SAME paddding in WinogradLayer by Pablo Tello · 6 years ago
  18. 4074c99 COMPMID-873: Integrate RSH NEON Depthwise Convolution routine by Georgios Pinitas · 6 years ago