4053ced2e597263ffe1c439bca8ca46e7b798cd2,gpytorch/lazy/lazy_evaluated_kernel_tensor.py,LazyEvaluatedKernelTensor,diag,#LazyEvaluatedKernelTensor#,152

Before Change

        if not self.is_batch:
            x1 = self.x1.unsqueeze(0)
            x2 = self.x2.unsqueeze(0)
        else:
            x1 = self.x1
            x2 = self.x2

After Change

        // TODO: until kernels support multi-batch mode, we have to remove any artificial batch dimension
        // that we"ve added
        if not is_batch:
            x1 = x1.squeeze(0)
            x2 = x2.squeeze(0)
            res = res.squeeze(0)

        // Did this Kernel eat the diag option?

Instances