quantized

Quantized Ops

This folder contains kernels for quantization related ops, similar to //executorch/kernels/portable/.

How to add quantized ops

1. Add a new operator definition in quantized.yaml

2. Implement the kernel for this operator.

3. Add unit test in /test directory.

4. Start to use //executorch/kernels/quantized:generated_lib in ExecuTorch or create your own generated lib if you only need a subset of the ops.