For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).
Mojo function
rms_norm_gpu_warp_per_row
def rms_norm_gpu_warp_per_row[mut: Bool, LayoutType: TensorLayout, origin: Origin[mut=mut], dtype: DType, Engine: TensorEngine, //, simd_width: Int, rows_per_block: Int, single_pass: Bool, chunks: Int, InputFnType: def[width: Int](Int, Int) -> SIMD[dtype, width] & RegisterPassable & ImplicitlyCopyable, OutputFnType: def[width: SIMDLength, alignment: Int](Int, Int, SIMD[dtype, width]) -> None & RegisterPassable & ImplicitlyCopyable, epsilon: Float32, weight_offset: Float32, multiply_before_cast: Bool, pdl_level: PDLLevel = PDLLevel.ON](gamma: TileTensor[dtype, LayoutType, origin, Engine=Engine], num_rows: Int32, num_cols: Int32, input_fn: InputFnType, output_fn: OutputFnType)