pytorch

f98bfce6 - Add const modifier for lazy shape inference for boolean logic ops (#79999)

Add const modifier for lazy shape inference for boolean logic ops (#79999) Follow up to https://github.com/pytorch/pytorch/pull/78004. --- Trying to fully codegen these boolean logic ops on pt/xla's side, saw errors: ``` Exception: Missing shape inference function. Please add declare this function in /workspace/pytorch/torch/csrc/lazy/core/shape_inference.h: and implement it in the the corresponding shape_inference.cpp file. TORCH_API std::vector<torch::lazy::Shape> compute_shape_logical_and(const at::Tensor & self, const at::Tensor & other); TORCH_API std::vector<torch::lazy::Shape> compute_shape_logical_not(const at::Tensor & self); TORCH_API std::vector<torch::lazy::Shape> compute_shape_logical_or(const at::Tensor & self, const at::Tensor & other); TORCH_API std::vector<torch::lazy::Shape> compute_shape_logical_xor(const at::Tensor & self, const at::Tensor & other); Failed to generate lazy files: ['python', '/workspace/pytorch/xla/scripts/gen_lazy_tensor.py'] ``` Add const modifier for lazy shape inference for boolean logic ops Pull Request resolved: https://github.com/pytorch/pytorch/pull/79999 Approved by: https://github.com/wconstab