misterBart

Shifted/reordered YOLO output layers when converting with TfLite 2.8 or higher

2

System info: Ubuntu 22.04 LTS Python: Python 3.10.4 Tensorflow Lite: 2.8 and higher Our YOLO network has three output layers. The (correct) output from Keras (on the .h5), TfLite2.5.0, and...

Thanks, questions, and suggestions

4

I thank you for providing FontPro, which helped me install Minion Pro and Myriad Pro. Still, a few questions and a few suggestions for improvement. Questions: (Q1) How can I...

Xnnpack still builds with `+dotprod` and `+fp16` with `-DXNNPACK_ENABLE_ARM_DOTPROD=OFF -DXNNPACK_ENABLE_ARM_FP16_SCALAR=OFF -DXNNPACK_ENABLE_ARM_FP16_VECTOR=OFF`

10

I'm building aan Arm64 target with a fairly old toolchain (gcc 7.5, binutils 2.29.1) in order to support old Linux platforms. I use: `-DXNNPACK_ENABLE_ARM_BF16=OFF -DXNNPACK_ENABLE_ARM_I8MM=OFF -DXNNPACK_ENABLE_ARM_DOTPROD=OFF -DXNNPACK_ENABLE_ARM_FP16_SCALAR=OFF -DXNNPACK_ENABLE_ARM_FP16_VECTOR=OFF ` Yet...