torch.js has not been released yet.

torch.nn.functional.ScaledMMFunctionalOptions

export interface ScaledMMFunctionalOptions {
  /** Scaling factor for the first matrix. (default: 1.0) */
  scale_a?: number;
  /** Scaling factor for the second matrix. (default: 1.0) */
  scale_b?: number;
  /** Scaling factor for the result. (default: 1.0) */
  scale_result?: number;
  /** Optional bias tensor to be added to the result. */
  bias?: Tensor;
  /** The desired data type of the returned tensor. (default: null) */
  out_dtype?: DType;
  /** If true, uses fast accumulation if supported by the hardware. (default: false) */
  use_fast_accum?: boolean;
}

scale_a(number)optional: – Scaling factor for the first matrix. (default: 1.0)
scale_b(number)optional: – Scaling factor for the second matrix. (default: 1.0)
scale_result(number)optional: – Scaling factor for the result. (default: 1.0)
bias(Tensor)optional: – Optional bias tensor to be added to the result.
out_dtype(DType)optional: – The desired data type of the returned tensor. (default: null)
use_fast_accum(boolean)optional: – If true, uses fast accumulation if supported by the hardware. (default: false)

Options for scaled_mm functional operation.

ScaledGroupedMMFunctionalOptions