{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":139209347,"defaultBranch":"master","name":"pytorch","ownerLogin":"malfet","currentUserCanPush":false,"isFork":true,"isEmpty":false,"createdAt":"2018-06-30T01:15:55.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/2453524?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1710864599.0","currentOid":""},"activityList":{"items":[{"before":null,"after":"94f30630c0aad04158c07cd47cdc11dde9513ec2","ref":"refs/heads/export-D55033347","pushedAt":"2024-03-19T16:09:59.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"[AARCH64] Hide FP16 scalar arithmetic behind proper feature flag\n\nTest Plan: CI\n\nReviewed By: dimitribouche\n\nDifferential Revision: D55033347","shortMessageHtmlLink":"[AARCH64] Hide FP16 scalar arithmetic behind proper feature flag"}},{"before":"2aa5c169cd61e696f11aca7649d3133c2227bda7","after":"b2449b6dc7da966d0931504796616f8adf780078","ref":"refs/heads/export-D54870507","pushedAt":"2024-03-13T22:22:15.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not compile FP16 arith internally (#121844)\n\nSummary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/121844\n\nTest Plan: `buck2 build -c fbcode.arch=aarch64  //caffe2:ATen-cpu`\n\nReviewed By: osalpekar\n\nDifferential Revision: D54870507","shortMessageHtmlLink":"Do not compile FP16 arith internally (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2184852411\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/121844\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/121844/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/121844\">pytorch#121844</a>)"}},{"before":"1311760c7cb6443bcdac2d1f39d167c24a4e5b0d","after":"2aa5c169cd61e696f11aca7649d3133c2227bda7","ref":"refs/heads/export-D54870507","pushedAt":"2024-03-13T21:11:39.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not compile FP16 arith internally (#121844)\n\nSummary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/121844\n\nTest Plan: `buck2 build -c fbcode.arch=aarch64  //caffe2:ATen-cpu`\n\nReviewed By: osalpekar\n\nDifferential Revision: D54870507","shortMessageHtmlLink":"Do not compile FP16 arith internally (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2184852411\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/121844\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/121844/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/121844\">pytorch#121844</a>)"}},{"before":"b4230f502633241563b301952450d0d6a8fc8463","after":"1311760c7cb6443bcdac2d1f39d167c24a4e5b0d","ref":"refs/heads/export-D54870507","pushedAt":"2024-03-13T21:02:27.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not compile FP16 arith internally (#121844)\n\nSummary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/121844\n\nTest Plan: `buck2 build -c fbcode.arch=aarch64  //caffe2:ATen-cpu`\n\nReviewed By: osalpekar\n\nDifferential Revision: D54870507","shortMessageHtmlLink":"Do not compile FP16 arith internally (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2184852411\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/121844\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/121844/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/121844\">pytorch#121844</a>)"}},{"before":null,"after":"b4230f502633241563b301952450d0d6a8fc8463","ref":"refs/heads/export-D54870507","pushedAt":"2024-03-13T20:43:14.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not compile FP16 arith internally\n\nTest Plan: `buck2 build -c fbcode.arch=aarch64  //caffe2:ATen-cpu`\n\nDifferential Revision: D54870507","shortMessageHtmlLink":"Do not compile FP16 arith internally"}},{"before":"0cfac48ab0be54ce5921a3a1778d66df3e798608","after":"bce56b916154e5a51a9b91657c1493752d10cdc9","ref":"refs/heads/export-D51358469","pushedAt":"2023-11-16T15:11:27.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not attempt to compile unwind.cpp on aarch64 (#113782)\n\nSummary:\n\nAs almost entire unwinding logic is build around x86_64 ABI\nIn essence, this reverts https://github.com/pytorch/pytorch/pull/104707 and adds `#ifndef FBCODE_CAFFE2` guards around `symbolize` dummy\n\nUse nested namespaces, as PyTorch is finally C++ compatible.\nRemove extraneous semicolon spotted by clang-tidy.\n\nFixes https://github.com/pytorch/pytorch/issues/113208\n\nTest Plan: CI + `buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file -c fbcode.arch=aarch64`\n\nReviewed By: aaronenyeshi\n\nDifferential Revision: D51358469","shortMessageHtmlLink":"Do not attempt to compile unwind.cpp on aarch64 (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1995295146\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/113782\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/113782/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/113782\">pytorch#113782</a>)"}},{"before":"b2bb6a570231e9b35a4315e614390085e840176e","after":"0cfac48ab0be54ce5921a3a1778d66df3e798608","ref":"refs/heads/export-D51358469","pushedAt":"2023-11-16T15:10:41.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not attempt to compile unwind.cpp on aarch64 (#113782)\n\nSummary:\n\nAs almost entire unwinding logic is build around x86_64 ABI\nIn essence, this reverts https://github.com/pytorch/pytorch/pull/104707 and adds `#ifndef FBCODE_CAFFE2` guards around `symbolize` dummy\n\nUse nested namespaces, as PyTorch is finally C++ compatible.\nRemove extraneous semicolon spotted by clang-tidy.\n\nFixes https://github.com/pytorch/pytorch/issues/113208\n\nTest Plan: CI + `buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file -c fbcode.arch=aarch64`\n\nReviewed By: aaronenyeshi\n\nDifferential Revision: D51358469","shortMessageHtmlLink":"Do not attempt to compile unwind.cpp on aarch64 (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1995295146\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/113782\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/113782/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/113782\">pytorch#113782</a>)"}},{"before":"c44cb585f2f60dda98e6591cf245422487a923fa","after":"b2bb6a570231e9b35a4315e614390085e840176e","ref":"refs/heads/export-D51358469","pushedAt":"2023-11-16T00:56:06.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not attempt to compile unwind.cpp on aarch64 (#113782)\n\nSummary:\n\nAs almost entire unwinding logic is build around x86_64 ABI\nIn essence, this reverts https://github.com/pytorch/pytorch/pull/104707 and adds `#ifndef FBCODE_CAFFE2` guards around `symbolize` dummy\n\nUse nested namespaces, as PyTorch is finally C++ compatible.\nRemove extraneous semicolon spotted by clang-tidy.\n\nFixes https://github.com/pytorch/pytorch/issues/113208\n\nTest Plan: CI + `buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file -c fbcode.arch=aarch64`\n\nReviewed By: aaronenyeshi\n\nDifferential Revision: D51358469","shortMessageHtmlLink":"Do not attempt to compile unwind.cpp on aarch64 (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1995295146\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/113782\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/113782/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/113782\">pytorch#113782</a>)"}},{"before":"a013c953fd29fa419f925020d06dd72e085ef04d","after":"c44cb585f2f60dda98e6591cf245422487a923fa","ref":"refs/heads/export-D51358469","pushedAt":"2023-11-16T00:49:41.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not attempt to compile unwind.cpp on aarch64 (#113782)\n\nSummary:\n\nAs almost entire unwinding logic is build around x86_64 ABI\nIn essence, this reverts https://github.com/pytorch/pytorch/pull/104707 and adds `#ifndef FBCODE_CAFFE2` guards around `symbolize` dummy\n\nUse nested namespaces, as PyTorch is finally C++ compatible.\nRemove extraneous semicolon spotted by clang-tidy.\n\nFixes https://github.com/pytorch/pytorch/issues/113208\n\nTest Plan: CI + `buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file -c fbcode.arch=aarch64`\n\nReviewed By: aaronenyeshi\n\nDifferential Revision: D51358469","shortMessageHtmlLink":"Do not attempt to compile unwind.cpp on aarch64 (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1995295146\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/113782\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/113782/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/113782\">pytorch#113782</a>)"}},{"before":"ebfe53f1c27703b86c13d4e5ea5be870ea6a5c5b","after":"a013c953fd29fa419f925020d06dd72e085ef04d","ref":"refs/heads/export-D51358469","pushedAt":"2023-11-16T00:37:27.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not attempt to compile unwind.cpp on aarch64 (#113782)\n\nSummary:\n\nAs almost entire unwinding logic is build around x86_64 ABI\nIn essence, this reverts https://github.com/pytorch/pytorch/pull/104707 and adds `#ifndef FBCODE_CAFFE2` guards around `symbolize` dummy\n\nAlso, use nested namespaces, as PyTorch is finally C++ compatible.\n\nFixes https://github.com/pytorch/pytorch/issues/113208\n\nTest Plan: CI + `buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file -c fbcode.arch=aarch64`\n\nReviewed By: aaronenyeshi\n\nDifferential Revision: D51358469","shortMessageHtmlLink":"Do not attempt to compile unwind.cpp on aarch64 (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1995295146\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/113782\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/113782/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/113782\">pytorch#113782</a>)"}},{"before":"3020c5797085ef6ea7d1119e8935efba6ffdf46c","after":"ebfe53f1c27703b86c13d4e5ea5be870ea6a5c5b","ref":"refs/heads/export-D51358469","pushedAt":"2023-11-16T00:37:11.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not attempt to compile unwind.cpp on aarch64 (#113782)\n\nSummary:\n\nAs almost entire unwinding logic is build around x86_64 ABI\nIn essence, this reverts https://github.com/pytorch/pytorch/pull/104707 and adds `#ifndef FBCODE_CAFFE2` guards around `symbolize` dummy\n\nFixes https://github.com/pytorch/pytorch/issues/113208\n\nTest Plan: CI + `buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file -c fbcode.arch=aarch64`\n\nReviewed By: aaronenyeshi\n\nDifferential Revision: D51358469","shortMessageHtmlLink":"Do not attempt to compile unwind.cpp on aarch64 (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1995295146\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/113782\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/113782/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/113782\">pytorch#113782</a>)"}},{"before":null,"after":"3020c5797085ef6ea7d1119e8935efba6ffdf46c","ref":"refs/heads/export-D51358469","pushedAt":"2023-11-15T18:18:01.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Do not attempt to compile unwind.cpp on aarch64\n\nSummary:\nAs almost entire unwinding logic is build around x86_64 ABI\nIn essence, this reverts https://github.com/pytorch/pytorch/pull/104707 and adds `#ifndef FBCODE_CAFFE2` guards around `symbolize` dummy\n\nFixes https://github.com/pytorch/pytorch/issues/113208\n\nTest Plan: CI + `buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file -c fbcode.arch=aarch64`\n\nDifferential Revision: D51358469","shortMessageHtmlLink":"Do not attempt to compile unwind.cpp on aarch64"}},{"before":"a0b2fbe87ee8259fa8706965747799968dec5e33","after":null,"ref":"refs/heads/export-D46161651","pushedAt":"2023-05-30T15:46:14.199Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"}},{"before":"21dfeafe6d63c3fd175ca02058994dd64d41035a","after":"a0b2fbe87ee8259fa8706965747799968dec5e33","ref":"refs/heads/export-D46161651","pushedAt":"2023-05-27T22:59:02.920Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Enable CuDNN v8 frontend in RL (#102284)\n\nSummary:\nThis enables use of CUDNN v8 in all Meta internal workflows. Also, fixes two minor issues:\n- Skip LogCumSumExp compilation for complex dtypes for fbcode and RL\n- Move `MakeConvOutputShape` template definition/specialization to anonymous namespace inside `at::native::quantized` as it is referenced from both `torch_cpu` and `torch_cuda`. This is necessary to avoid `duplicate symbol` linker error if say `libtorch_cpu` and `libtorch_cuda` are statically linked together.\n- Lower CuDNN v8 version guard from 8.3 to 8.2 (as there are no good reason why it should be 8.3, first version of the library that properly supports all the features is actually 8.5)\n\nPull Request resolved: https://github.com/pytorch/pytorch/pull/102284\n\nTest Plan: buck2 build arvr/mode/platform010/cuda11/opt //xplat/caffe2:_C\n\nReviewed By: atalman\n\nDifferential Revision: D46161651\n\nPulled By: malfet\n\nfbshipit-source-id: 796a11951efcef6da7901806f091fc3927ee4867","shortMessageHtmlLink":"Enable CuDNN v8 frontend in RL (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1726251099\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/102284\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/102284/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/102284\">pytorch#102284</a>)"}},{"before":"1083fe4bf0e4283c6f7692459331f7c929e39261","after":"21dfeafe6d63c3fd175ca02058994dd64d41035a","ref":"refs/heads/export-D46161651","pushedAt":"2023-05-27T22:43:08.005Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Enable CuDNN v8 frontend in RL (#102284)\n\nSummary:\nThis enables use of CUDNN v8 in all Meta internal workflows. Also, fixes two minor issues:\n- Skip LogCumSumExp compilation for complex dtypes for fbcode and RL\n- Move `MakeConvOutputShape` template definition/specialization to anonymous namespace inside `at::native::quantized` as it is referenced from both `torch_cpu` and `torch_cuda`. This is necessary to avoid `duplicate symbol` linker error if say `libtorch_cpu` and `libtorch_cuda` are statically linked together.\n- Lower CuDNN v8 version guard from 8.3 to 8.2 (as there are no good reason why it should be 8.3, first version of the library that properly supports all the features is actually 8.5)\n\nPull Request resolved: https://github.com/pytorch/pytorch/pull/102284\n\nTest Plan: buck2 build arvr/mode/platform010/cuda11/opt //xplat/caffe2:_C\n\nReviewed By: atalman\n\nDifferential Revision: D46161651\n\nPulled By: malfet\n\nfbshipit-source-id: dffdd70f16ac0ce163ebc4344b11f27f991e9b97","shortMessageHtmlLink":"Enable CuDNN v8 frontend in RL (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1726251099\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/102284\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/102284/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/102284\">pytorch#102284</a>)"}},{"before":"6e93a4abbd02aef754ba123703bd248a9aa7f4f0","after":"1083fe4bf0e4283c6f7692459331f7c929e39261","ref":"refs/heads/export-D46161651","pushedAt":"2023-05-27T22:36:18.351Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Enable CuDNN v8 frontend in RL (#102284)\n\nSummary:\nThis enables use of CUDNN v8 in all Meta internal workflows. Also, fixes two minor issues:\n- Skip LogCumSumExp compilation for complex dtypes for fbcode and RL\n- Move `MakeConvOutputShape` template definition/specialization to anonymous namespace inside `at::native::quantized` as it is referenced from both `torch_cpu` and `torch_cuda`. This is necessary to avoid `duplicate symbol` linker error if say `libtorch_cpu` and `libtorch_cuda` are statically linked together.\n- Lower CuDNN v8 version guard from 8.3 to 8.2 (as there are no good reason why it should be 8.3, first version of the library that properly supports all the features is actually 8.5)\n\nPull Request resolved: https://github.com/pytorch/pytorch/pull/102284\n\nTest Plan: buck2 build arvr/mode/platform010/cuda11/opt //xplat/caffe2:_C\n\nReviewed By: atalman\n\nDifferential Revision: D46161651\n\nPulled By: malfet\n\nfbshipit-source-id: dcb0ebe0becf360f5567fbe76e90b6914e4731b2","shortMessageHtmlLink":"Enable CuDNN v8 frontend in RL (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1726251099\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/102284\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/102284/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/102284\">pytorch#102284</a>)"}},{"before":"5f3ced729906bb202b2b26c2e597408f9845fd3a","after":"6e93a4abbd02aef754ba123703bd248a9aa7f4f0","ref":"refs/heads/export-D46161651","pushedAt":"2023-05-27T22:19:32.700Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Move to namespace and add TORCH=API","shortMessageHtmlLink":"Move to namespace and add TORCH=API"}},{"before":"ae3bcd48bba1721919d0b504f37302984c3d2d6d","after":"5f3ced729906bb202b2b26c2e597408f9845fd3a","ref":"refs/heads/export-D46161651","pushedAt":"2023-05-25T21:37:19.199Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Enable CuDNN v8 frontend in RL (#102284)\n\nSummary:\nThis enables use of CUDNN v8 in all Meta internal workflows. Also, fixes two minor issues:\n- Skip LogCumSumExp compilation for complex dtypes for fbcode and RL\n- Move `MakeConvOutputShape` template defiinition/specialization to a common location, as technically one should not have two template specialization in the .cpp files.\n\nPull Request resolved: https://github.com/pytorch/pytorch/pull/102284\n\nTest Plan:\nCI\n\ncc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10\n\nReviewed By: atalman\n\nDifferential Revision: D46161651\n\nPulled By: malfet\n\nfbshipit-source-id: 66d869471221de3ba08fbaf2727858db73692255","shortMessageHtmlLink":"Enable CuDNN v8 frontend in RL (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"1726251099\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/102284\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/102284/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/102284\">pytorch#102284</a>)"}},{"before":"b9e9150c6e8e6632d0a9ab9f8c96d3bc0b66ed03","after":"ae3bcd48bba1721919d0b504f37302984c3d2d6d","ref":"refs/heads/export-D46161651","pushedAt":"2023-05-25T21:17:29.577Z","pushType":"push","commitsCount":1,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Move to namespace and add TORCH=API","shortMessageHtmlLink":"Move to namespace and add TORCH=API"}},{"before":null,"after":"b9e9150c6e8e6632d0a9ab9f8c96d3bc0b66ed03","ref":"refs/heads/export-D46161651","pushedAt":"2023-05-25T17:56:16.044Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Enable CuDNN v8 frontend in RL\n\nSummary:\nThis enables use of CUDNN v8 in all Meta internal workflows. Also, fixes two minor issues:\n- Skip LogCumSumExp compilation for complex dtypes for fbcode and RL\n- Move `MakeConvOutputShape` template defiinition/specialization to a common location, as technically one should not have two template specialization in the .cpp files.\n\nTest Plan: CI\n\nDifferential Revision: D46161651\n\nfbshipit-source-id: 93e854fe4c6700215e4863e30ff5a31cede75d92","shortMessageHtmlLink":"Enable CuDNN v8 frontend in RL"}},{"before":null,"after":"2a567f3434375d5c639dcf6f558d164f8cc8fa90","ref":"refs/heads/export-D45027286","pushedAt":"2023-04-17T17:52:09.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"Revert D44897935: Multisect successfully blamed D44897935 for test or build failures\n\nSummary:\nThis diff is reverting D44897935\nD44897935: [FSDP] Include duplicate parameters and modules when calling named_parameters and named_modules (#98912) by fegin has been identified to be causing the following test or build failures:\n\nTests affected:\n- [caffe2/torch/fb/module_factory/sync_sgd/tests:test_pyper_data_parallel_wrapper - caffe2.torch.fb.module_factory.sync_sgd.tests.test_pyper_data_parallel_wrapper.PyPerDataParallelWrapperTest: test_fsdp_submodules_pyper](https://www.internalfb.com/intern/test/562950025957458/)\n\nHere's the Multisect link:\nhttps://www.internalfb.com/multisect/1893714\nHere are the tasks that are relevant to this breakage:\n\nWe're generating a revert to back out the changes in this diff, please note the backout may land if someone accepts it.\n\nIf you believe this diff has been generated in error you may Commandeer and Abandon it.\n\nTest Plan: NA\n\nReviewed By: fegin\n\nDifferential Revision: D45027286\n\nfbshipit-source-id: ba19640b24371140886a6eca2c6dbfa52cfa77f2","shortMessageHtmlLink":"Revert D44897935: Multisect successfully blamed D44897935 for test or…"}},{"before":null,"after":"a9fa438a3d5958e330bd6b2bbe14254df3175d72","ref":"refs/heads/malfet/be-remove-unused-dtype","pushedAt":"2023-03-30T20:13:09.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"malfet","name":"Nikita Shulga","path":"/malfet","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/2453524?s=80&v=4"},"commit":{"message":"[BE] Remove bool dtype from `masked_scatter`","shortMessageHtmlLink":"[BE] Remove bool dtype from <code>masked_scatter</code>"}}],"hasNextPage":false,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEGdBo7QA","startCursor":null,"endCursor":null}},"title":"Activity · malfet/pytorch"}