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_quantize_mxfp8_q_kv_kernel
def rms_norm_quantize_mxfp8_q_kv_kernel[output_layout: TensorLayout, scales_layout: TensorLayout, kv_layout: TensorLayout, input_layout: TensorLayout, q_gamma_layout: TensorLayout, kv_gamma_layout: TensorLayout, *, q_cols: Int, kv_cols: Int](output: TileTensor[.float8_e4m3fn, output_layout, MutAnyOrigin], scales: TileTensor[.float8_e8m0fnu, scales_layout, MutAnyOrigin], kv_output: TileTensor[.bfloat16, kv_layout, MutAnyOrigin], input: TileTensor[.bfloat16, input_layout, ImmutAnyOrigin], q_gamma: TileTensor[.bfloat16, q_gamma_layout, ImmutAnyOrigin], kv_gamma: TileTensor[.bfloat16, kv_gamma_layout, ImmutAnyOrigin], epsilon: Float32, num_rows: Int32)