Topic · Edition
Auto-generated nightly by .github/workflows/nightly.yml. Updates the pinned triton hash to the tip of release/3.8.x. Pull Request resolved: #192864 Approved by: https://github.com/pytorchbot
Not node index. process_subgraph_nodes walks the subgraph with enumerate() and uses that index to read args: for i, node in enumerate(graph_module.graph.nodes): if node.op == "placeholder": V.graph.env[node] = args[i]…
[ROCm] Enable FP8 FNUZ sparse semi-structured support on ROCm MI300+ …
Revert "Increase dynamo coverage of the operator module by adding p…
Changes since langchain-openai==1.4.3 release(openai): 1.5.0 (#39629) feat(openai): support openai 3.0 SDK (#39613) chore(partners): bump langgraph floor in openai and huggingface lockfiles (#39617)
This reverts commit 9bd9dc6. Reverted #179286 on behalf of https://github.com/atalman due to breaks nightly (comment)
This PR is auto-generated nightly by this action. Update the pinned vllm hash. Pull Request resolved: #193245 Approved by: https://github.com/pytorchbot
This PR is auto-generated nightly by this action. Update the pinned vllm hash. Pull Request resolved: #193245 Approved by: https://github.com/pytorchbot
This PR is auto-generated nightly by this action. Update the pinned torchcomms hash. Pull Request resolved: #193246 Approved by: https://github.com/pytorchbot
Enable eager tests for inductor-unittest - define xfails for xpu in o…
Fail the job when a compile-worker result callback dies (#192301) (#1…
Fail the job when a compile-worker result callback dies (#192301) (#1…
C++20 provides a contains member function for all associative containers. Prior to C++20, the common pattern to check if an element is a member of a container was to either container.find(key) != container.end() or…
We have C++20 in PyTorch now #176662 Spaceship/Three-way comparison operator automatically generates <, <=, >, >= operator== automatically generates operator!= This reduces a little bit of code. Pull Request resolved:…
Changes since langchain-anthropic==1.5.5 release(anthropic): 1.5.6 (#39622) fix(anthropic): normalize tool_search_tool_result blocks (#39621) fix(anthropic): correct model profile data for Fable 5, Sonnet 5, Opus 4.1…
Update the torch-xpu-ops commit to intel/torch-xpu-ops@9061f2: Add SYCL-based FFT kernel as an alternative to oneMKL FFT Fix TopK AOT compilation issues in subgroup and single-workgroup kernels Trim oneMKL LAPACK…
Motivation As title Additional Test passed Pull Request resolved: #191182 Approved by: https://github.com/EikanWang
Motivation As title Additional Test passed Pull Request resolved: #191182 Approved by: https://github.com/EikanWang
Bump cuDNN to 9.25.0.15 for CUDA 13.4rc1 wheels (#192843) cuDNN 9.24.0.43 has no fused attention engines for compute capability 10.7, so scaled_dot_product_attention fails on Rubin hardware with RuntimeError: cuDNN…
Retry and validate the randomtemp.exe download on Windows CUDA builds…
Fall back to 'triage review' instead of blocking when all labels are …
Fix silently wrong second-order forward AD for linalg.det/slogdet/sol…
