# PyTorch v2.14.0 — PyTorch 2.14.0 Release - Product: PyTorch (https://whatsnew.fyi/product/pytorch) - Vendor: PyTorch Foundation - Date: 2026-09-02 - Version: v2.14.0 - Original notes: https://github.com/pytorch/pytorch/releases/tag/v2.14.0 - Permalink: https://whatsnew.fyi/product/pytorch/releases/v2.14.0 What's New is an index, not a publisher: every entry below links to the vendor's own release notes, which are the authoritative source. Entries are labelled where they are hand-curated sample data, pre-releases, or drawn from a secondary source such as a developer blog. Reuse: the summaries, labels and curation here are © What's New. Quote freely with attribution and a link back; wholesale republication of the corpus is not permitted — terms: https://whatsnew.fyi/terms. The vendors' own release notes remain their publishers'. --- - **removed** — torch.nn.LinearCrossEntropyOptions no longer accepts acc_policy="balanced"; use "compact" instead - **changed** — Clamp and min/max boundary subgradients now follow the selected dispatcher schema's input space, changing gradient behavior at nondifferentiable bounds or ties - **changed** — Custom Python process groups that implement new_group() must now accept a backend keyword argument - **added** — NVGEMM brings CuTeDSL-generated CUTLASS kernels to Inductor with epilogue fusion, scaled and NVFP4 GEMM, and grouped-reduction epilogues - **added** — torch.switch generalizes torch.cond to multi-way branching - **added** — torch.while_loop can now be captured in a CUDA graph - **added** — Declarative dynamic shapes via @dynamic_spec shared across torch.compile, torch.export and make_fx - **added** — Experimental torch.compile support for complex-valued tensors with opt-in deposition of complex operations into real and imaginary computations - **added** — New nccl2 backend for PyTorch Distributed with nonblocking communicators and eager communicator splitting - **added** — Fault tolerance as a first-class c10d concept with in-place process-group reconfiguration, one-sided RMA windows, and Flight Recorder - **added** — Apple Silicon gains native linear algebra including Jacobi-kernel SVD, eigh, QR and Cholesky - **added** — ROCm 7.14 wheels are produced from the TheRock pip SDK - **added** — Intel XPU adds native graph capture - **added** — Inductor targets Rubin (sm_107) #### PyTorch 2.14.0 Release Notes - [Highlights](#highlights) - [Backwards Incompatible Changes](#backwards-incompatible-changes) - [Deprecations](#deprecations) - [New Features](#new-features) - [Improvements](#improvements) - [Bug fixes](#bug-fixes) - [Performance](#performance) - [Documentation](#documentation) - [Security](#security) - [Developers](#developers) #### Highlights
NVGEMM brings CuTeDSL-generated CUTLASS kernels to Inductor, with epilogue fusion, scaled and NVFP4 GEMM, and grouped-reduction epilogues autotuned alongside Triton and ATen
torch.switch generalizes torch.cond to multi-way branching, and torch.while_loop can now be captured in a CUDA graph
Declarative dynamic shapes via @dynamic_spec, shared across torch.compile, torch.export and make_fx
Experimental torch.compile support for complex-valued tensors: Opt-in support decomposes supported complex operations into real and imaginary computations, enabling compiler backends to optimize more complex-number workloads.
A new nccl2 backend for PyTorch Distributed, ported from torchcomms, implementing the full collective contract with nonblocking communicators and eager communicator splitting
Fault tolerance becomes a first-class c10d concept, with in-place process-group reconfiguration, one-sided RMA windows, and a Flight Recorder that works for any backend rather than only NCCL
Apple Silicon gains native linear algebra, including Jacobi-kernel SVD, eigh, QR and Cholesky, alongside a five-part reduction rewrite and a further MPSGraph to Metal kernel migration
Broader platform support: ROCm 7.14 wheels are produced from the TheRock pip SDK, Intel XPU adds native graph capture, and Inductor targets Rubin (sm_107)
For more details about these highlighted features, you can look at the release blogpost. Below are the full release notes for this release. #### Backwards Incompatible Changes ##### torch.nn - `torch.nn.LinearCrossEntropyOptions` no longer accepts `acc_policy="balanced"`; use `"compact"` instead (#188283) The `"balanced"` policy was removed because `"compact"` provides the same weight-gradient accumulation precision with lower memory use on CUDA, already uses the equivalent scratch layout for mixed-precision inputs on other devices, and was never selected by `"auto"`. Constructing the options with `acc_policy="balanced"` now raises `ValueError: invalid acc_policy: 'balanced'; expected one of 'auto', 'accurate', 'compact'`. Before: ```python options = torch.nn.LinearCrossEntropyOptions(acc_policy="balanced") loss = torch.nn.functional.linear_cross_entropy( input, linear_weight, target, options=options ) ``` After: ```python options = torch.nn.LinearCrossEntropyOptions(acc_policy="compact") loss = torch.nn.functional.linear_cross_entropy( input, linear_weight, target, options=options ) ``` ##### Autograd - Clamp and min/max boundary subgradients now follow the selected dispatcher schema's input space (#191142) This affects gradients exactly at nondifferentiable bounds or ties. A scalar clamp bound is a fixed parameter, so the input gradient at equality changes from `1` to the minimum-norm subgradient `0`. A Tensor bound is part of the differentiable input space, so `clamp`, `clamp_min`, and `clamp_max` now split the gradient evenly between the input and bound at an ordinary tie instead of assigning it entirely to the input. `fmin` and `fmax` use the same even tie split, and forward-mod _[Truncated at 4000 characters — full notes: https://github.com/pytorch/pytorch/releases/tag/v2.14.0]_