@thi.ng/tensors/softmax.d.ts

0D/1D/2D/3D/4D tensors with extensible polymorphic operations and customizable storage

import type { ITensor } from "./api.js";
/**
 * Computes softmax (aka normalized {@link exp}) of input tensor `src` and
 * writes results to `out` (or if null, back into `src`). The optional
 * `temperature` (should be >= 1) can be used to adjust the emphasis/scaling of
 * normalized values: Higher temperatures reduce the contrast ratio between
 * min/max components.
 *
 * @remarks
 * Computes: `smax = exp(src / temp) / sum(exp(src / temp))`. The result
 * vector's components will all be in the [0,1] range and sum to 1.0.
 *
 * This function is often used as the last activation function in a neural
 * network to normalize the output to a probability distribution over predicted
 * output classes.
 *
 * References:
 *
 * - https://en.wikipedia.org/wiki/Softmax_function
 * - https://victorzhou.com/blog/softmax/
 *
 * @param out -
 * @param src -
 * @param temperature -
 */
export declare const softMax: (out: ITensor | null, src: ITensor, temperature?: number) => import("./tensor.js").Tensor1<number>;
//# sourceMappingURL=softmax.d.ts.map