Building PyTorch from source with CUDA versions older than 12.6 is no longer supported. Users building custom binaries should install CUDA 12.6 or newer and make sure `CUDA_HOME` points to that installation.
Deprecations
Deprecate CUDA 12.8 builds in favor of CUDA 13.0 ([#179072](https://github.com/pytorch/pytorch/pull/179072))
Compatibility with CMake < 3.10 will be removed in a future release ([#166259](https://github.com/pytorch/pytorch/pull/166259))
Source builds against CMake versions older than 3.10 now emit a deprecation warning. A future release will require CMake 3.10 or newer; please upgrade CMake before then.
Several CUDA linear algebra operators no longer use the MAGMA backend and now dispatch to cuSolver or cuBLAS unconditionally:
Volta (SM 7.0) GPU support removed from CUDA 12.8 and 12.9 binary builds (#172598)
CUDA 12.8 builds no longer support Volta
torch.hub.list()`, `torch.hub.load()`, and `torch.hub.help()` now default the `trust_repo` parameter to `"check"` instead of `None`. The `trust_repo=None` option has been removed. (#174101)
Remove `is_causal` flag from `varlen_attn` (#172245)
DTensor.to_local()` backward now converts `Partial` placements to `Replicate` by default when `grad_placements` is not provided. (#173454)
Users who relied on the previous behavior (where `to_local()` backward preserved `Partial` gradients) may see different gradient values. To ensure correctness, explicitly pass `grad_placements` to `to_local()`.
Partial placement preserved in backward — could produce incorrect gradients
Remove Nested Jagged Tensor support from `nn.attention.flex_attention` ([#161734](https://github.com/pytorch/pytorch/pull/161734))
Deprecations
We decided to deprecate an existing behavior which goes against the PyTorch design principle (explicit over implicit) for device mesh slicing of flattened dim.
torch.jit` is not guaranteed to work in Python 3.14. Deprecation warnings have been added to user-facing `torch.jit` API ([#167669](https://github.com/pytorch/pytorch/pull/167669)).
The `dynamic_axes` option in `torch.onnx.export` is deprecated ([#165769](https://github.com/pytorch/pytorch/pull/165769))
<td>The ability to arbitrarily toggle error or resume on graph breaks in torch.compile</td
Backwards Incompatible Changes
Build metal kernels of MacOS-14+ and remove all pre-MacOS-14 specific logic, requires MacOS-14+ going forward (#159733, #159912)
This upgrade is doing the same BC-breaking changes as the DLPack release. Objects in `torch.utils.dlpack` have been updated to reflect these changes, such as `DLDeviceType`.
Remove implicit draft tracing from the default exporter path, achieving clearer behaviour and faster failures.
The expensive `torch.export.draft_export` diagnostic path is no longer auto-invoked (which could take hours on large models). You can still opt in for deep diagnostics:
Remove `torch.onnx.dynamo_export` and the `onnxrt` torch compile backend (#158130, #158258)
torch.onnx.dynamo_export` is removed. Please use `torch.onnx.export` instead.