{"full_name":"arogozhnikov/einops","owner":"arogozhnikov","name":"einops","shard":"1","url":"https://github.com/arogozhnikov/einops","description":"Flexible and powerful tensor operations for readable and reliable code (for pytorch, jax, TF and others)","homepage":"https://einops.rocks","topics":["deep-learning","pytorch","tensorflow","numpy","cupy","tensor","jax","einops","mlx"],"tag_pages":["deep-learning","jax","mlx","numpy","pytorch","tensor","tensorflow"],"language":"Python","license":"MIT","avatar":"https://avatars.githubusercontent.com/u/6318811?u=5e99e15d9400a4bd8a449939093d38df94fd3e90&v=4","created_at":"2018-09-22T00:45:08Z","pushed_at":"2026-08-26T08:23:21Z","metrics":{"stars":9637,"forks":409,"watchers":60,"issues":207,"prs":201,"releases":17,"issues_open":31,"issues_closed":176,"prs_open":16,"prs_closed":25,"prs_merged":160,"discussions":34},"growth":{"day":{"stars":1,"forks":0,"base_day":"2026-11-15"}},"history":{"dates":["2026-11-10","2026-11-11","2026-11-12","2026-11-13","2026-11-14","2026-11-15","2026-11-16"],"stars":[9633,9633,9634,9634,9635,9636,9637],"forks":[407,408,408,408,409,409,409],"watchers":[null,60,60,60,60,60,60],"issues_open":[null,31,31,31,31,31,31],"prs_open":[null,15,15,15,16,16,16]},"boards":{},"same_owner":[],"issues":[{"number":84,"title":"Generic einops function which infers which operation to call","comments":18,"created_at":"2020-10-27T01:01:52Z"},{"number":22,"title":"Ellipsis not mentioned in docs","comments":14,"created_at":"2018-12-01T10:52:24Z"},{"number":194,"title":"[Feature suggestion] einops for array indexing","comments":12,"created_at":"2022-06-23T20:59:49Z"},{"number":185,"title":"Missing einops.layers.....Repeat","comments":6,"created_at":"2022-04-26T21:25:04Z"},{"number":168,"title":"[Feature suggestion] Add a simple shape assertion in einops-style notation","comments":6,"created_at":"2022-01-21T20:20:01Z"},{"number":88,"title":"do einops' operations account for contiguous memory layout?","comments":5,"created_at":"2020-11-16T12:53:19Z"},{"number":107,"title":"einops and tensorboard graphs","comments":4,"created_at":"2021-04-07T14:05:04Z"},{"number":20,"title":"Why differently shaped tensors can't be concatenated with rearrange?","comments":4,"created_at":"2018-11-20T02:43:54Z"}],"discussions":[{"number":404,"title":"einops.rearrange optimization","comments":1,"category":"Ideas","created_at":"2026-02-21T06:59:42Z"},{"number":380,"title":"argmin, argmax, argwhere","comments":2,"category":"Q&A","created_at":"2025-07-15T16:33:00Z"},{"number":311,"title":"Improving the expressivity of einsum with lambda operators","comments":2,"category":"Ideas","created_at":"2024-03-06T15:54:49Z"},{"number":310,"title":"Optimizing order of operations","comments":1,"category":"Ideas","created_at":"2024-03-04T05:55:32Z"},{"number":307,"title":"Is it possible to look at `numpy` representation of concrete `einops` function?","comments":1,"category":"Ideas","created_at":"2024-02-05T23:02:52Z"},{"number":301,"title":"Masking in Einops","comments":1,"category":"Ideas","created_at":"2024-01-10T10:50:03Z"},{"number":291,"title":"Assign to diagonal","comments":2,"category":"Ideas","created_at":"2023-11-06T21:11:48Z"},{"number":280,"title":"Something like np.outer","comments":1,"category":"Q&A","created_at":"2023-09-15T13:37:22Z"}],"latest_release":{"name":"v0.8.2: full MLX backend, rely on torch.compile, set python 3.9 as minimum","tag":"v0.8.2","published_at":"2026-01-26T04:12:47Z","body":"\r\n\r\n- torch.compile: previously we relied on registration of einops in torch.compile. This registration was called by torch. Now with torch >= 2.8 `torch.compile` engine can compile einops operations without additional hints/registrations.\r\n- recommended backend for MLX is still `einops.array_api.<operation>`, but since community asked multiple times for support of `einops.einsum`, there is now MLX backend in einops, so `einops.<operation>` would also work with MLX.\r\n- minor improvements to docs, and mention of several related projects.\r\n\r\n\r\n\r\n\r\n\r\n\r\n## What's Changed\r\n* Update README.md by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/366\r\n* Update codebase, introduce consistency checks, update EinMix tutorial by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/374\r\n* Update python versions in tests, change min supported version to 3.9 by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/364\r\n* mention cubed as one more backend via array api by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/375\r\n* Extend tests for torch.compile by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/382\r\n* add einops.jl and einops for R in docs_src/pages/projects.md by @Qile0317 in https://github.com/arogozhnikov/einops/pull/383\r\n* mention ein color and AlphaFold3 by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/389\r\n* A bit more of compile tests by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/390\r\n* Patch #391 (conflicts in multithreading initialization) by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/392\r\n* switch off debug importing by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/393\r\n* Skip `allow_in_graph` if torch >= 2.8 by @guilhermeleobas in https://github.com/arogozhnikov/einops/pull/394\r\n* Add mlx backend by @ifsheldon in https://github.com/arogozhnikov/einops/pull/376\r\n* update testing by @arogozhnikov in https://github.com/arogozhnikov/einops/pull/395\r\n\r\n## New Contributors\r\n* @Qile0317 made their first contribution in https://github.com/arogozhnikov/einops/pull/383\r\n* @guilhermeleobas made their first contribution in https://github.com/arogozhnikov/einops/pull/394\r\n* @ifsheldon made their first contribution in https://github.com/arogozhnikov/einops/pull/376\r\n\r\n**Full Changelog**: https://github.com/arogozhnikov/einops/compare/v0.8.1...v0.8.2","is_prerelease":false},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-16T09:23:14Z","history_start":"2026-09-06"}