# Uncategorized **URL:** https://dev-discuss.pytorch.org/c/uncategorized/1.md?page=1 [Latest](https://dev-discuss.pytorch.org/latest.md) · [Categories](https://dev-discuss.pytorch.org/categories.md) **Page:** 2 --- ## [Questions about how to contribute to pytorch](https://dev-discuss.pytorch.org/t/questions-about-how-to-contribute-to-pytorch/3255)
**Author:** [@wenlinchong17-web](https://dev-discuss.pytorch.org/u/wenlinchong17-web)\ **Replies:** 6\ **Last updated:** [December 12, 2025, 5:39am UTC](https://dev-discuss.pytorch.org/t/questions-about-how-to-contribute-to-pytorch/3255 "2025-12-12T05:39:23Z")
Hello, I am a third-year computer science student currently enrolled in a course titled Open Source Software Development. I have chosen PyTorch as my project. This is my first time contributing to such a project, and I w… --- ## [S3 Access denied for ROCm nightlies](https://dev-discuss.pytorch.org/t/s3-access-denied-for-rocm-nightlies/3277)
**Author:** [@ewindisch](https://dev-discuss.pytorch.org/u/ewindisch)\ **Replies:** 0\ **Last updated:** [December 8, 2025, 7:05pm UTC](https://dev-discuss.pytorch.org/t/s3-access-denied-for-rocm-nightlies/3277 "2025-12-08T19:05:01Z")
curl -i https://download.pytorch.org/libtorch/nightly/rocm7.1/libtorch-shared-with-deps-latest.zip HTTP/2 403 content-type: application/xml server: AmazonS3 date: Mon, 08 Dec 2025 19:00:46 GMT x-cache: Error from cl… --- ## [Testing in PrivateUse1 for out-of-tree PyTorch Backends](https://dev-discuss.pytorch.org/t/testing-in-privateuse1-for-out-of-tree-pytorch-backends/3270)
**Author:** [@albanD](https://dev-discuss.pytorch.org/u/albanD)\ **Replies:** 6\ **Last updated:** [December 3, 2025, 8:56pm UTC](https://dev-discuss.pytorch.org/t/testing-in-privateuse1-for-out-of-tree-pytorch-backends/3270 "2025-12-03T20:56:46Z")
Testing in PrivateUse1 and out of core backend has been quickly evolving and so I wanted to write down a quick update here on where we are today and where we expect to go short term. See the new User Guide for details o… --- ## [Torch.export with aot-indutor for cute dsl kernels](https://dev-discuss.pytorch.org/t/torch-export-with-aot-indutor-for-cute-dsl-kernels/3272)
**Author:** [@jfmontgar](https://dev-discuss.pytorch.org/u/jfmontgar)\ **Replies:** 3\ **Last updated:** [December 1, 2025, 3:24pm UTC](https://dev-discuss.pytorch.org/t/torch-export-with-aot-indutor-for-cute-dsl-kernels/3272 "2025-12-01T15:24:57Z")
Hi, I was wondering if there is a mechanism to export a model with torch.export that contains a torch.library.custom\_op that calls a CUDA kernel. Ideally, I would like to use a python dsl (like triton or cute dsl) if… --- ## [PyTorch Core Maintainers Meeting - Dec 5th 2025](https://dev-discuss.pytorch.org/t/pytorch-core-maintainers-meeting-dec-5th-2025/3269)
**Author:** [@gottbrath](https://dev-discuss.pytorch.org/u/gottbrath)\ **Replies:** 0\ **Last updated:** [November 18, 2025, 5:28pm UTC](https://dev-discuss.pytorch.org/t/pytorch-core-maintainers-meeting-dec-5th-2025/3269 "2025-11-18T17:28:48Z")
Our next PyTorch Core Maintainer Meeting is scheduled for Dec 5th. The general purpose of these meetings is to: Discuss any technical plans and roadmaps for PyTorch Review any high-level requests related to PyTorc… --- ## [Meta PyTorch Team 2025 H2 Roadmaps](https://dev-discuss.pytorch.org/t/meta-pytorch-team-2025-h2-roadmaps/3184)
**Author:** [@gottbrath](https://dev-discuss.pytorch.org/u/gottbrath)\ **Replies:** 6\ **Last updated:** [August 28, 2025, 8:21pm UTC](https://dev-discuss.pytorch.org/t/meta-pytorch-team-2025-h2-roadmaps/3184 "2025-08-28T20:21:06Z")
PyTorch Community, The Meta team are happy to make our 2025 H2 roadmaps available. We plan on a half year basis and globally optimize across the things we do for our users here at Meta and across the wider community. We… --- ## [\[RFC\] Graph Neural Network Library for LibTorch - Seeking Community Input](https://dev-discuss.pytorch.org/t/rfc-graph-neural-network-library-for-libtorch-seeking-community-input/3221)
**Author:** [@devesht21](https://dev-discuss.pytorch.org/u/devesht21)\ **Replies:** 0\ **Last updated:** [August 27, 2025, 6:19pm UTC](https://dev-discuss.pytorch.org/t/rfc-graph-neural-network-library-for-libtorch-seeking-community-input/3221 "2025-08-27T18:19:26Z")
Hi PyTorch Community! :waving\_hand: ## The Problem I’ve noticed a significant gap in the PyTorch ecosystem: \*\*there’s no comprehensive Graph Neural Network (GNN) library for LibTorch/C++\*\*. While Python has excellent … --- ## [Can't find implementation of nll\_loss\_nd](https://dev-discuss.pytorch.org/t/cant-find-implementation-of-nll-loss-nd/3149)
**Author:** [@PydevC](https://dev-discuss.pytorch.org/u/PydevC)\ **Replies:** 1\ **Last updated:** [August 22, 2025, 10:18am UTC](https://dev-discuss.pytorch.org/t/cant-find-implementation-of-nll-loss-nd/3149 "2025-08-22T10:18:16Z")
I wanted to checkout the for implementation of torch.nn.functional.nll\_loss() I came across this line where nll\_loss was implemented. Looking into the code at the location torch.\_c.\_nn.pyi I was unable to find code rela… --- ## [How to get started with open-source contributing](https://dev-discuss.pytorch.org/t/how-to-get-started-with-open-source-contributing/3219)
**Author:** [@TheRecklessDoctor](https://dev-discuss.pytorch.org/u/TheRecklessDoctor)\ **Replies:** 0\ **Last updated:** [August 22, 2025, 9:14am UTC](https://dev-discuss.pytorch.org/t/how-to-get-started-with-open-source-contributing/3219 "2025-08-22T09:14:07Z")
Hi I am a researcher looking to get started with open-source contributions to pytorch. How do I best explore the codebase so that if and when I start working on a issue, I won’t get too confused and can find my way about… --- ## [How to install MPS backend on docker/podman container?](https://dev-discuss.pytorch.org/t/how-to-install-mps-backend-on-docker-podman-container/3015)
**Author:** [@pytorch1](https://dev-discuss.pytorch.org/u/pytorch1)\ **Replies:** 2\ **Last updated:** [August 18, 2025, 10:08am UTC](https://dev-discuss.pytorch.org/t/how-to-install-mps-backend-on-docker-podman-container/3015 "2025-08-18T10:08:41Z")
I’m trying to install pytorch with MPS backend - has anyone tried installing it on Docker? --- ## [New Contributor Looking for Mentorship on a pull request!](https://dev-discuss.pytorch.org/t/new-contributor-looking-for-mentorship-on-a-pull-request/3179)
**Author:** [@vulkomilev](https://dev-discuss.pytorch.org/u/vulkomilev)\ **Replies:** 0\ **Last updated:** [August 13, 2025, 4:50pm UTC](https://dev-discuss.pytorch.org/t/new-contributor-looking-for-mentorship-on-a-pull-request/3179 "2025-08-13T16:50:14Z")
Hi my name is Valko Milev and who’s passionate about contributing to PyTorch! But I need some guidance . I am currently fully employed and I am doing some ml projects on the side and unfortunately I don’t have enough t… --- ## [Understanding the difference between the caching behavior of cuda caching allocator and pluggable allocator](https://dev-discuss.pytorch.org/t/understanding-the-difference-between-the-caching-behavior-of-cuda-caching-allocator-and-pluggable-allocator/2746)
**Author:** [@youkaichao](https://dev-discuss.pytorch.org/u/youkaichao)\ **Replies:** 7\ **Last updated:** [August 2, 2025, 11:47pm UTC](https://dev-discuss.pytorch.org/t/understanding-the-difference-between-the-caching-behavior-of-cuda-caching-allocator-and-pluggable-allocator/2746 "2025-08-02T23:47:11Z")
Recently, I’m playing around PyTorch’s memory allocator, and I’d like to record here what I found. First, as we all know, PyTorch uses cuda caching allocator by default. To monitor all allocation and free from the drive… --- ## [Meta PyTorch Team 2025 H1 Roadmaps](https://dev-discuss.pytorch.org/t/meta-pytorch-team-2025-h1-roadmaps/2794)
**Author:** [@gottbrath](https://dev-discuss.pytorch.org/u/gottbrath)\ **Replies:** 17\ **Last updated:** [June 24, 2025, 11:19pm UTC](https://dev-discuss.pytorch.org/t/meta-pytorch-team-2025-h1-roadmaps/2794 "2025-06-24T23:19:08Z")
PyTorch Community, The Meta team are happy to make our 2025 H1 roadmaps available. We plan on a half year basis and globally optimize across the things we do for our users here at Meta and across the wider community. We… --- ## [Does pytorch's CI have a queue-jumping function?](https://dev-discuss.pytorch.org/t/does-pytorchs-ci-have-a-queue-jumping-function/3063)
**Author:** [@swgu98](https://dev-discuss.pytorch.org/u/swgu98)\ **Replies:** 4\ **Last updated:** [June 24, 2025, 10:43pm UTC](https://dev-discuss.pytorch.org/t/does-pytorchs-ci-have-a-queue-jumping-function/3063 "2025-06-24T22:43:32Z")
Does pytorch’s CI have a queue-jumping function? Does GitHub action have a queue-jumping feature? --- ## [Can I get access to the PyTorch Slack channel?](https://dev-discuss.pytorch.org/t/can-i-get-access-to-the-pytorch-slack-channel/3071)
**Author:** [@narekmalk](https://dev-discuss.pytorch.org/u/narekmalk)\ **Replies:** 0\ **Last updated:** [June 17, 2025, 5:54pm UTC](https://dev-discuss.pytorch.org/t/can-i-get-access-to-the-pytorch-slack-channel/3071 "2025-06-17T17:54:29Z")
I have authored these PRs --- ## [How to register a custom operator with multiple in-place operations in PyTorch?](https://dev-discuss.pytorch.org/t/how-to-register-a-custom-operator-with-multiple-in-place-operations-in-pytorch/3045)
**Author:** [@ff\_pytorch](https://dev-discuss.pytorch.org/u/ff_pytorch)\ **Replies:** 1\ **Last updated:** [June 6, 2025, 6:53am UTC](https://dev-discuss.pytorch.org/t/how-to-register-a-custom-operator-with-multiple-in-place-operations-in-pytorch/3045 "2025-06-06T06:53:28Z")
I’m trying to register an operator that contains multiple in-place operations, defined as: - func: custom\_demo\_add\_(Tensor(a!) input1, Tensor(b!) input2) → (Tensor(a!), Tensor(b!)) But I’m getting an error: ‘If you ha… --- ## [How to add custom attributes to pytorch tensor](https://dev-discuss.pytorch.org/t/how-to-add-custom-attributes-to-pytorch-tensor/3018)
**Author:** [@bailuan](https://dev-discuss.pytorch.org/u/bailuan)\ **Replies:** 3\ **Last updated:** [May 28, 2025, 2:34am UTC](https://dev-discuss.pytorch.org/t/how-to-add-custom-attributes-to-pytorch-tensor/3018 "2025-05-28T02:34:03Z")
How can I add custom attributes like device\_local or host\_local to a PyTorch tensor without affecting TensorImpl or StorageImpl? I have a use case where I need to convert an external tensor into a PyTorch tensor while pr… --- ## [Tensor\_totype issues with device doesn't support double](https://dev-discuss.pytorch.org/t/tensor-totype-issues-with-device-doesnt-support-double/2947)
**Author:** [@chn-lee-yumi](https://dev-discuss.pytorch.org/u/chn-lee-yumi)\ **Replies:** 7\ **Last updated:** [May 27, 2025, 3:34am UTC](https://dev-discuss.pytorch.org/t/tensor-totype-issues-with-device-doesnt-support-double/2947 "2025-05-27T03:34:43Z")
When the device doesn’t support double (eg. intel iGPUs with torch-directml), this function will throw RuntimError: def tensor\_totype(t): dtype = torch.float if t.is\_mps else torch.double return t.to(dtype=dtype… --- ## [Just Started Contributing](https://dev-discuss.pytorch.org/t/just-started-contributing/3024)
**Author:** [@Achintyap](https://dev-discuss.pytorch.org/u/Achintyap)\ **Replies:** 0\ **Last updated:** [May 23, 2025, 10:45pm UTC](https://dev-discuss.pytorch.org/t/just-started-contributing/3024 "2025-05-23T22:45:34Z")
Hi there! I have made 2 PRs to the PyTorch repository and was wondering how I can get involved in the community itself. Are there meetings for all contributors? How do we get better to eventually become a Issue Triage o… --- ## [Fix to make torch.load safer in Pytorch v2.1.2](https://dev-discuss.pytorch.org/t/fix-to-make-torch-load-safer-in-pytorch-v2-1-2/2992)
**Author:** [@Aman-Surkar](https://dev-discuss.pytorch.org/u/Aman-Surkar)\ **Replies:** 1\ **Last updated:** [May 21, 2025, 8:57pm UTC](https://dev-discuss.pytorch.org/t/fix-to-make-torch-load-safer-in-pytorch-v2-1-2/2992 "2025-05-21T20:57:46Z")
I am trying to fix CVE-2025-32434 in Pytorch v2.1.2, as I won’t be able to upgrade Pytorch to v2.6.0. I tried backporting PRs (#137602, #138225, #138866, #139221, #140304, #138936, #139541, #140738, #142153, #139433) men… --- ## [Balance "Usability over Performance" in a concentrating ML Ecosystem](https://dev-discuss.pytorch.org/t/balance-usability-over-performance-in-a-concentrating-ml-ecosystem/2930)
**Author:** [@albanD](https://dev-discuss.pytorch.org/u/albanD)\ **Replies:** 3\ **Last updated:** [May 20, 2025, 9:22pm UTC](https://dev-discuss.pytorch.org/t/balance-usability-over-performance-in-a-concentrating-ml-ecosystem/2930 "2025-05-20T21:22:11Z")
The very first item in the PyTorch Design Philosophy is “Usability over Performance”. As discussed in the design page, while this principle can be surprising, it is at the core of PyTorch flexibility and has been key in… --- ## [Run lintrunner locally on Windows machine](https://dev-discuss.pytorch.org/t/run-lintrunner-locally-on-windows-machine/3000)
**Author:** [@narekmalk](https://dev-discuss.pytorch.org/u/narekmalk)\ **Replies:** 4\ **Last updated:** [May 18, 2025, 4:19pm UTC](https://dev-discuss.pytorch.org/t/run-lintrunner-locally-on-windows-machine/3000 "2025-05-18T16:19:06Z")
When running lintrunner init, I am getting this error: ERROR: Unsupported platform: Windows/Windows-Intel64 --- ## [Which papers are relevant to understand how the core of PyTorch and the GPU kernels and torch.compile are put together?](https://dev-discuss.pytorch.org/t/which-papers-are-relevant-to-understand-how-the-core-of-pytorch-and-the-gpu-kernels-and-torch-compile-are-put-together/2983)
**Author:** [@Pawl](https://dev-discuss.pytorch.org/u/Pawl)\ **Replies:** 6\ **Last updated:** [May 16, 2025, 7:06pm UTC](https://dev-discuss.pytorch.org/t/which-papers-are-relevant-to-understand-how-the-core-of-pytorch-and-the-gpu-kernels-and-torch-compile-are-put-together/2983 "2025-05-16T19:06:05Z")
Hi, I’m a newbie who is embarking on this project of studying PyTorch holistically. At this point I am mostly interested in the core of PyTorch, the CUDA kernels and the torch.compile project. I am already somewhat famil… --- ## [Backporting CVE fixes from Pytorch v2.6.0 to v2.1.2](https://dev-discuss.pytorch.org/t/backporting-cve-fixes-from-pytorch-v2-6-0-to-v2-1-2/2990)
**Author:** [@Aman-Surkar](https://dev-discuss.pytorch.org/u/Aman-Surkar)\ **Replies:** 0\ **Last updated:** [May 12, 2025, 2:26pm UTC](https://dev-discuss.pytorch.org/t/backporting-cve-fixes-from-pytorch-v2-6-0-to-v2-1-2/2990 "2025-05-12T14:26:08Z")
As per the discussion on issue Backport CVEs in Pytorch v2.1.2 from Pytorch v2.6.0 · Issue #153370 · pytorch/pytorch, is Prevent torch.jit.load path in torch.load when weights\_only=True by mikaylagawarecki · Pull Request… --- ## [Clarification of PyTorch Quantization Flow Support (in pytorch and torchao)](https://dev-discuss.pytorch.org/t/clarification-of-pytorch-quantization-flow-support-in-pytorch-and-torchao/2809)
**Author:** [@jerryzh168](https://dev-discuss.pytorch.org/u/jerryzh168)\ **Replies:** 9\ **Last updated:** [May 9, 2025, 8:41pm UTC](https://dev-discuss.pytorch.org/t/clarification-of-pytorch-quantization-flow-support-in-pytorch-and-torchao/2809 "2025-05-09T20:41:48Z")
There has been some questions around the support for torchao quantization flow(quantize\_, autoquant, QAT etc.) and pt2e quantization flow(prepare\_pt2e, prepare\_qat\_pt2e, convert\_pt2e) by PyTorch. This note is trying to g… --- ## [New Contributor Looking for Mentorship!](https://dev-discuss.pytorch.org/t/new-contributor-looking-for-mentorship/2637)
**Author:** [@EmmettBicker](https://dev-discuss.pytorch.org/u/EmmettBicker)\ **Replies:** 5\ **Last updated:** [April 17, 2025, 9:30pm UTC](https://dev-discuss.pytorch.org/t/new-contributor-looking-for-mentorship/2637 "2025-04-17T21:30:33Z")
Hi everyone, I’m Emmett, an eighteen-year-old who’s passionate about contributing to PyTorch! I’ve been self-teaching myself machine learning for the past three years and I took a gap year before entering college to do … --- ## [CUDA Kernel with customised device](https://dev-discuss.pytorch.org/t/cuda-kernel-with-customised-device/2918)
**Author:** [@Ishu2001](https://dev-discuss.pytorch.org/u/Ishu2001)\ **Replies:** 1\ **Last updated:** [April 17, 2025, 1:18pm UTC](https://dev-discuss.pytorch.org/t/cuda-kernel-with-customised-device/2918 "2025-04-17T13:18:27Z")
is there way to launch cuda kernels with the customised device, useprivate1… or to customise cuda and create new device on top of it, by o verloading some operations --- ## [Float8 in PyTorch \[1/x\]](https://dev-discuss.pytorch.org/t/float8-in-pytorch-1-x/1815)
**Author:** [@vkuzo](https://dev-discuss.pytorch.org/u/vkuzo)\ **Replies:** 1\ **Last updated:** [April 7, 2025, 7:02pm UTC](https://dev-discuss.pytorch.org/t/float8-in-pytorch-1-x/1815 "2025-04-07T19:02:16Z")
This note introduces the experimental support for float8 training in PyTorch. The float8 data formats (as laid out in https://arxiv.org/pdf/2209.05433.pdf) have been added in recent hardware such as NVIDIA Hopper. They h… --- ## [Rename prune.remove](https://dev-discuss.pytorch.org/t/rename-prune-remove/2891)
**Author:** [@Laubeee](https://dev-discuss.pytorch.org/u/Laubeee)\ **Replies:** 0\ **Last updated:** [April 7, 2025, 3:49pm UTC](https://dev-discuss.pytorch.org/t/rename-prune-remove/2891 "2025-04-07T15:49:44Z")
The name remove is highly confusing, and IMO should be renamed. Some examples could be: finalize / finalize\_pruning commit / commit\_pruning consolidate / consolidate\_pruning persist bake another option could be make\_p… --- ## [Higher Order Operators, 2023/10](https://dev-discuss.pytorch.org/t/higher-order-operators-2023-10/1565)
**Author:** [@zou3519](https://dev-discuss.pytorch.org/u/zou3519)\ **Replies:** 2\ **Last updated:** [March 25, 2025, 2:37am UTC](https://dev-discuss.pytorch.org/t/higher-order-operators-2023-10/1565 "2025-03-25T02:37:03Z")
Higher Order Operators, 2023/10 What do autograd.Function, torch.utils.checkpoint, torch.cond, triton kernels, and torch.vmap all have in common? Answer: torch.compile support for the above APIs (and more!) go through th… [Previous page](https://dev-discuss.pytorch.org/c/uncategorized/1.md) [Next page](https://dev-discuss.pytorch.org/c/uncategorized/1.md?page=2)