2491af9de0dd8d20c4bb8236d9c51cdb33f14060,deepexplain/tensorflow/methods.py,IntegratedGradients,run,#IntegratedGradients#,172

Before Change

        self.baseline = baseline

    def run(self):
        if self.baseline is None: self.baseline = np.zeros((1,)+self.xs.shape[1:])
        elif self.baseline.shape == self.xs.shape[1:]:
            self.baseline = np.expand_dims(self.baseline, 0)
        else:
            raise RuntimeError("Baseline shape %s does not match expected shape %s"
                               % (self.baseline.shape, self.xs.shape[1:]))
        attributions = self.get_symbolic_attribution()
        gradient = None
        for alpha in list(np.linspace(1. / self.steps, 1.0, self.steps)):
            xs_mod = self.xs * alpha

After Change

            xs_mod = (np.array(self.xs) * alpha).tolist()
            _attr = self.session_run(attributions, xs_mod)
            if gradient is None: gradient = np.array(_attr)
            else: gradient += np.array(_attr)
        results = gradient * (np.array(self.xs) - np.array(self.baseline)) / self.steps
        return results[0] if not self.has_multiple_inputs else results

Instances