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, Storage: TensorStorage, //, simd_width: Int, rows_per_block: Int, single_pass: Bool, chunks: Int, input_fn: def[width: Int](row: Int, col: Int) capturing thin -> SIMD[dtype, width], output_fn: def[width: SIMDLength, alignment: Int](row: Int, col: Int, val: SIMD[dtype, width]) capturing thin -> None, multiply_before_cast: Bool](gamma: TileTensor[dtype, LayoutType, origin, Storage=Storage], epsilon: Float32, weight_offset: Float32, num_rows: Int32, num_cols: Int32)