• edited by pytorch-bot bot 🐛 Describe the bug Code to reproduce: import torch device = torch.device('cpu') dtype = torch.bfloat16 n = 512 x = torch.tensor([0, 1], dtype=dtype, device=device).repeat(n // 2) y = torch.zeros(n, device=device, dtype=torch.int32) print(...
TypeError: allreduce(): incompatible function arguments. The following argument types are supported: 1. (self: torch.distributed.ProcessGroup, arg0: List[at::Tensor], arg1: torch.distributed.AllreduceOptions) -> c10d::ProcessGroup::Work 2. (self: torch.distributed.ProcessGroup, tensor: at::Te...
We've received your issue and please be patient to get responded. We will arrange technicians to...
Assign User on Comment DISABLED test_inductor_reduce_scatter_tensor_single (__main__.CompileTest) #153042 Sign in to view logs Summary Jobs assign Run details Usage Workflow file Triggered via issue February 26, 2025 06:43 pytorch-bot[bot] commented on #147911 4216478 Status Success Total...
Disabled by pytorch-bot[bot] Within ~15 minutes, test_reduce_scatter_tensor_device_cpu_use_native_funcol_False (__main__.TestTraceableCollectives) will be disabled in PyTorch CI for these platforms: linux. Please verify that your test name looks correct, e.g., test_cuda_assert_async (__...
pytorch-bot bot added ciflow/inductor module: dynamo module: inductor labels Jun 13, 2024 jamesjwu mentioned this pull request Jun 13, 2024 Ignore functional tensor wrapper when caching #128335 Closed jamesjwu added a commit that referenced this pull request Jun 13, 2024 Reduce all tensor...
UserWarning: The reduce argument of torch.scatter with Tensor src is deprecated and will be removed in a future PyTorch release. Use torch.scatter_reduce instead for more reduction options. in test_scatter_add and test_scatter_mul. In case you wonder why I didn't replace every usage of to...
Please update the following actions to use Node.js 20: pytorch/test-infra/.github/actions/setup-ssh@main, malfet/checkout@silent-checkout, nick-fields/retry@3e91a01664abd3c5cd539100d10d33b9c5b68482, seemethere/upload-artifact-s3@v5. For more information see: https://github.blog/changelog...
pytorchmergebot added the Merged label Jun 4, 2024 bigfootjon pushed a commit that referenced this pull request Jun 5, 2024 [dtensor][debug] added c10d reduce_scatter_ and reduce_scatter_tensor… … 88f5015 bigfootjon pushed a commit that referenced this pull request Jun 5, 2024 ...
Collaborator ngimel commented Mar 27, 2025 • edited by pytorch-bot bot Per title, we want to be able to use it even if inputs are not registered. Separate copy would add latency, and one-shot is all about the lowest possible latency. cc @H-Huang @awgu @kwen2501 @wanchaol @feg...