recurrent-js-gpu/dist/R.js

GPU-accelerated Deep Recurrent Neural Networks and LSTMs in Typescript. Ported, object-oriented and refactored version of Andrej Karpathy's recurrent-js (https://github.com/karpathy/recurrentjs)

"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
const Assertable_1 = require("./utils/Assertable");
class R extends Assertable_1.Assertable {
    static randf(min, max) { return Math.random() * (max - min) + min; }
    static randi(min, max) { return Math.floor(R.randf(min, max)); }
    static randn(mu, std) { return mu + R.gaussRandom() * std; }
    static gaussRandom() {
        if (R.returnV) {
            R.returnV = false;
            return R.vVal;
        }
        const u = 2 * Math.random() - 1;
        const v = 2 * Math.random() - 1;
        const r = u * u + v * v;
        if (r === 0 || r > 1) {
            return R.gaussRandom();
        }
        const c = Math.sqrt(-2 * Math.log(r) / r);
        R.vVal = v * c;
        R.returnV = true;
        return u * c;
    }
    static fillRandn(m, mu, std) { for (let i = 0; i < m.w.length; i++) {
        m.w[i] = R.randn(mu, std);
    } }
    static fillRand(m, lo, hi) { for (let i = 0; i < m.w.length; i++) {
        m.w[i] = R.randf(lo, hi);
    } }
    static gradFillConst(m, c) { for (let i = 0; i < m.dw.length; i++) {
        m.dw[i] = c;
    } }
    static setConst(arr, c) {
        for (let i = 0; i < arr.length; i++) {
            arr[i] = c;
        }
    }
    static zeros(n) {
        if (typeof (n) === 'undefined' || isNaN(n)) {
            return [];
        }
        if (typeof ArrayBuffer === 'undefined') {
            const arr = new Array(n);
            R.setConst(arr, 0);
            return arr;
        }
        else {
            return new Float64Array(n);
        }
    }
    static maxi(w) {
        let maxv = w[0];
        let maxix = 0;
        for (let i = 1; i < w.length; i++) {
            const v = w[i];
            if (v > maxv) {
                maxix = i;
                maxv = v;
            }
        }
        return maxix;
    }
    static sampleWeighted(p) {
        const r = Math.random();
        let c = 0.0;
        for (let i = 0; i < p.length; i++) {
            c += p[i];
            if (c >= r) {
                return i;
            }
        }
        this.assert(false, 'wtf');
    }
}
R.returnV = false;
R.vVal = 0.0;
exports.R = R;
//# sourceMappingURL=R.js.map