44826d877d755e7df1d681d680583911d6d2032e,gpytorch/kernels/kernel.py,Distance,_jit_sq_dist_x1_neq_x2_nobatch,#Distance#Any#Any#Any#,37

Before Change

    // @torch.jit.script_method
    def _jit_sq_dist_x1_neq_x2_nobatch(self, x1, x2, postprocess):
        // Compute squared distance matrix using quadratic expansion
        x1_norm = x1.pow(2).sum(dim=-1, keepdim=True)
        x2_norm = x2.pow(2).sum(dim=-1, keepdim=True)

        res = torch.addmm(x2_norm.transpose(-2, -1), x1, x2.transpose(-2, -1), alpha=-2).add_(x1_norm)

        // Zero out negative values
        res.clamp_min_(0)
        return self._postprocess(res) if bool(postprocess) else res

    // @torch.jit.script_method
    def _jit_sq_dist_x1_eq_x2(self, x1, postprocess):

After Change

    // @torch.jit.script_method
    def _jit_sq_dist_x1_neq_x2_nobatch(self, x1, x2, postprocess):
        if hasattr(torch, "cdist"):  // TODO: Remove lower branch when PyTorch 1.1.0 releases.
            res = torch.cdist(x1, x2).pow(2)
        else:
            // Compute squared distance matrix using quadratic expansion
            x1_norm = x1.pow(2).sum(dim=-1, keepdim=True)
            x2_norm = x2.pow(2).sum(dim=-1, keepdim=True)

            res = torch.addmm(x2_norm.transpose(-2, -1), x1, x2.transpose(-2, -1), alpha=-2).add_(x1_norm)

            // Zero out negative values
            res.clamp_min_(0)
        return self._postprocess(res) if bool(postprocess) else res

    // @torch.jit.script_method
    def _jit_sq_dist_x1_eq_x2(self, x1, postprocess):

Instances