[ExecuTorch][WebGPU] rms_norm: add a vec4 kernel for 4-aligned row widths#20458

meta-codesync[bot] merged 4 commits into

gh/JulianCloudNTH/53/basepytorch/executorch:gh/JulianCloudNTH/53/basefrom

gh/JulianCloudNTH/53/headpytorch/executorch:gh/JulianCloudNTH/53/headCopy head branch name to clipboard