pytorch
pytorch copied to clipboard
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Summary: co-dev reland of https://github.com/pytorch/pytorch/pull/124520, which requires the removal of some executorch tests. Before this PR, we didn't check that types in a schema were valid. This is because TorchScript...
Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #126823 Fix an issue reported by internal user that conv lowering does not work well with dynamic shapes. cc @voznesenskym @penguinwu @EikanWang...
Summary: To check what fake class we need to add to make all tests pass. Test Plan: Existing tests. Differential Revision: D57678693
### 🐛 Describe the bug Currently, PyTorch crashes if two functions annotated with `@functools.wraps(func)` are composed together. As an example, when adding support for `torch.func.hessian` in https://github.com/pytorch/pytorch/pull/121410, `wraps` is conditionally...
#fix https://github.com/pytorch/pytorch/issues/126355 Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #126904 * #126693 * __->__ #126444 cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang
Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #126841 **Summary** Fix issue: https://github.com/pytorch/pytorch/issues/126824 which is missing the support of `min/max` with `VecMask`. **TestPlan** ``` python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_max_cpu_bool python test/inductor/test_torchinductor_opinfo.py...
Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #126872 We were calling backward on a tensor not a scalar...
Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #126893 * __->__ #126876 cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @peterbell10 @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames...
This PR is intended to provide the Autograd mechanism for the `cond` operation. @ydwu4