bef7dd92cbdd2c81a7d89786b9a76b011f1bc086,maml_rl/baseline.py,LinearFeatureBaseline,fit,#LinearFeatureBaseline#Any#,29

Before Change

                torch.matmul(featmat.t(), returns),
                torch.matmul(featmat.t(), featmat) + reg_coeff * eye
            )
            if not isnan(self.linear.weight).any():
                break
            reg_coeff *= 10
        self.linear.weight.data = coeffs.data.t()

    def forward(self, episodes):

After Change

        reg_coeff = self._reg_coeff
        eye = torch.eye(self.feature_size, dtype=torch.float32,
            device=self.linear.weight.device)
        for _ in range(5):
            try:
                coeffs, _ = torch.gels(
                    torch.matmul(featmat.t(), returns),
                    torch.matmul(featmat.t(), featmat) + reg_coeff * eye
                )
                break
            except RuntimeError:
                reg_coeff += 10
        else:
            raise RuntimeError("Unable to solve the normal equations in "
                "`LinearFeatureBaseline`. The matrix X^T*X (with X the design "
                "matrix) is not full-rank, regardless of the regularization "
                "(maximum regularization: {0}).".format(reg_coeff))
        self.linear.weight.copy_(coeffs.t())

    def forward(self, episodes):
        features = self._feature(episodes)

Instances