def _step(self, inp): self.outs = [] for i in range(self.max_iter): self.s = np.sum(self.np['w'] * inp, axis=1) self.s += self.np['b'] out = self.transf(self.s) if i > 0 and np.abs(out - inp).sum() <= self.delta: break self.outs.append(out) inp = out return out
def initnw(layer): """ Nguyen-Widrow initialization function """ ci = layer.ci cn = layer.cn w_fix = 0.7 * cn ** (1. / ci) w_rand = np.random.rand(cn, ci) * 2 - 1 # Normalize if ci == 1: w_rand = w_rand / np.abs(w_rand) else: w_rand = w_rand * np.sqrt(1. / np.square(w_rand).sum(axis=1).reshape(cn, 1)) w = w_fix * w_rand b = np.array([0]) if cn == 1 else w_fix * np.linspace(-1, 1, cn) * np.sign(w[:, 0]) # Scaleble to inp_active amin, amax = layer.transf.inp_active amin = -1 if amin == -np.Inf else amin amax = 1 if amax == np.Inf else amax x = 0.5 * (amax - amin) y = 0.5 * (amax + amin) w = x * w b = x * b + y # Scaleble to inp_minmax minmax = layer.inp_minmax.copy() minmax[np.isneginf(minmax)] = -1 minmax[np.isinf(minmax)] = 1 x = 2. / (minmax[:, 1] - minmax[:, 0]) y = 1. - minmax[:, 1] * x w = w * x b += np.dot(w, y) layer.np['w'][:] = w layer.np['b'][:] = b return
def __call__(self, e): v = np.sum(np.abs(e)) / e.size return v
def __call__(self, e): v = np.sum(np.abs(e)) return v