{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":335514537,"defaultBranch":"master","name":"pytorch","ownerLogin":"mdja","currentUserCanPush":false,"isFork":true,"isEmpty":false,"createdAt":"2021-02-03T05:12:19.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/604124?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1618261865.2608562","currentOid":""},"activityList":{"items":[{"before":"444203c52fe116a988bca8c71ab101aeabb88335","after":"ca98c2a932132e49559bf777c02798633d585e66","ref":"refs/heads/master","pushedAt":"2024-05-03T13:01:58.000Z","pushType":"push","commitsCount":10000,"pusher":{"login":"mdja","name":"Muhammad Djamaluddin","path":"/mdja","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/604124?s=80&v=4"},"commit":{"message":"inductor: Add Conv3d support (#124361)\n\nThis PR is to add Conv3d support in inductor. Basicly reuse and expand Conv2d logic and unit tests to Conv3d.\n\nConv3d inductor support will improve the performance of C2D_R50, I3D_R50, I3D_R101, Slow and SlowFast-R50 from OOB models.\n\n  | C2D_R50 | I3D_R50 | I3D_R101 | Slow | SlowFast-R50\n-- | -- | -- | -- | -- | --\neager | 15.805 | 13.909 | 11.639 | 12.101 | 6.606\nCompile w/o conv3d | 17.244 | 14.893 | 12.109 | 13.015 | 6.603\nCompile w/ conv3d | 21.212 | 17.707 | 14.974 | 16.130 | 8.537\n\nPull Request resolved: https://github.com/pytorch/pytorch/pull/124361\nApproved by: https://github.com/leslie-fang-intel, https://github.com/CaoE, https://github.com/jgong5, https://github.com/jansel","shortMessageHtmlLink":"inductor: Add Conv3d support (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2249793676\" data-permission-text=\"Title is private\" data-url=\"https://github.com/pytorch/pytorch/issues/124361\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/pytorch/pytorch/pull/124361/hovercard\" href=\"https://github.com/pytorch/pytorch/pull/124361\">pytorch#124361</a>)"}}],"hasNextPage":false,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEQMN7CgA","startCursor":null,"endCursor":null}},"title":"Activity · mdja/pytorch"}