{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":756234502,"defaultBranch":"main","name":"optimum-tpu","ownerLogin":"huggingface","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2024-02-12T08:54:38.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/25720743?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1726760766.0","currentOid":""},"activityList":{"items":[{"before":null,"after":"93ca2601e6cdf49594ea33b8ebddcb0e499dea2f","ref":"refs/heads/quick-fixes-for-jetstream","pushedAt":"2024-09-19T15:46:06.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"feat(Jetstream Pt): set torch default dtype to bfloat16\n\nThis is what should be used on TPUs.","shortMessageHtmlLink":"feat(Jetstream Pt): set torch default dtype to bfloat16"}},{"before":"23ab23b421f2d891f96936cc415eb01c1a06f77a","after":"8519c8187392116a96b61dc1a2082f7810617e73","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-19T14:57:02.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"refactor(Jetstream Pt): avoid duplicating Llama modeling\n\nSince this is error-prone, a better solution is just to use this.\nThis hadn't been done before mainly because in the model config we do\nnot have some of the params anymore (ffn_dim_multiplier and\nmultiple_of). We do have intermediate_size though, and that is enough to\nreconstruct parameters that end up producing the same calculation.\n\nThis refactor should allow for future code to follow Jetstream/Pytorch\nchanges in an easier way.","shortMessageHtmlLink":"refactor(Jetstream Pt): avoid duplicating Llama modeling"}},{"before":"fa50f3b32ec84928812ba4ff3f456a046c5c7275","after":"23ab23b421f2d891f96936cc415eb01c1a06f77a","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-19T14:23:45.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"refactor(Jetstream Pt): avoid duplicating Llama modeling\n\nSince this is error-prone, a better solution is just to use this.\nThis hadn't been done before mainly because in the model config we do\nnot have some of the params anymore (ffn_dim_multiplier and\nmultiple_of). We do have intermediate_size though, and that is enough to\nreconstruct parameters that end up producing the same calculation.\n\nThis refactor should allow for future code to follow Jetstream/Pytorch\nchanges in an easier way.","shortMessageHtmlLink":"refactor(Jetstream Pt): avoid duplicating Llama modeling"}},{"before":"a8cad0af35c1d6755b74e7ddfebac8b390e3e0c0","after":"fa50f3b32ec84928812ba4ff3f456a046c5c7275","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-19T14:20:48.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"refactor(Jetstream Pt): avoid duplicating Llama modeling\n\nSince this is error-prone, a better solution is just to use this.\nThis hadn't been done before mainly because in the model config we do\nnot have some of the params anymore (ffn_dim_multiplier and\nmultiple_of). We do have intermediate_size though, and that is enough to\nreconstruct parameters that end up producing the same calculation.\n\nThis refactor should allow for future code to follow Jetstream/Pytorch\nchanges in an easier way.","shortMessageHtmlLink":"refactor(Jetstream Pt): avoid duplicating Llama modeling"}},{"before":"edb7ee36171d63c695bde5ba8b4992d21c44d0cc","after":"a8cad0af35c1d6755b74e7ddfebac8b390e3e0c0","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-18T15:43:39.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"fix(jetstream Pt): make Jetstream Pt install more reliable\n\nI was previously referencing a given git revision and install from\ngithub, but since the Jetstream Pytorch package install its dependencies\nfrom its git submodels, these are installed in temporary directories,\nthat can disappear afterwards. This happened on CI, making the\ninstallation fail.\n\nTo work around that, a dedicated install script has been added, and it\nis now used to install that.","shortMessageHtmlLink":"fix(jetstream Pt): make Jetstream Pt install more reliable"}},{"before":"b0e6a1ac7d2ed9b96fd4f5f80f8869f6a8ed7425","after":"edb7ee36171d63c695bde5ba8b4992d21c44d0cc","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-18T15:40:26.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"fix(jetstream Pt): make Jetstream Pt install more reliable\n\nI was previously referencing a given git revision and install from\ngithub, but since the Jetstream Pytorch package install its dependencies\nfrom its git submodels, these are installed in temporary directories,\nthat can disappear afterwards. This happened on CI, making the\ninstallation fail.\n\nTo work around that, a dedicated install script has been added, and it\nis now used to install that.","shortMessageHtmlLink":"fix(jetstream Pt): make Jetstream Pt install more reliable"}},{"before":"d7b897883e2bfd9a2238327636e0f2f42c20f56a","after":"b0e6a1ac7d2ed9b96fd4f5f80f8869f6a8ed7425","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-18T13:37:36.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP: disable main test pipeline","shortMessageHtmlLink":"WIP: disable main test pipeline"}},{"before":"6ec35ed3cebf94df8083330286bbd9d11dc687c2","after":"d7b897883e2bfd9a2238327636e0f2f42c20f56a","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-18T12:42:39.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"fix(Jetstream Pt): correct generation\n\nText generation was not correct because the weights in the model were\nnot correctly loaded. This is not something that it was easy to spot\njust looking at few tokens generated, and it was something that it was\nactually fixed already in the Jetstream/Pytorch code, but the fix hadn't\nbeen ported to optimum-tpu.\n\nThis fix implement the necessary weights changes, aligning to Jetstream\nPytorch, and tests expected output has been modified accordingly.","shortMessageHtmlLink":"fix(Jetstream Pt): correct generation"}},{"before":"28da563144d6b318ebed580c2e191dc111ab6958","after":"6ec35ed3cebf94df8083330286bbd9d11dc687c2","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-18T12:42:08.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"fix(Jetstream Pt): correct generation\n\nText generation was not correct because the weights in the model were\nnot correctly loaded. This is not something that it was easy to spot\njust looking at few tokens generated, and it was something that it was\nactually fixed already in the Jetstream/Pytorch code, but the fix hadn't\nbeen ported to optimum-tpu.\n\nThis fix implement the necessary weights changes, aligning to Jetstream\nPytorch, and tests expected output has been modified accordingly.","shortMessageHtmlLink":"fix(Jetstream Pt): correct generation"}},{"before":"7b9c0ca6610ce68bd470e230456c3d1ed435e5b3","after":"28da563144d6b318ebed580c2e191dc111ab6958","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-18T11:42:22.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"fix(Jetstream Pt): correct generation\n\nText generation was not correct because the weights in the model were\nnot correctly loaded. This is not something that it was easy to spot\njust looking at few tokens generated, and it was something that it was\nactually fixed already in the Jetstream/Pytorch code, but the fix hadn't\nbeen ported to optimum-tpu.\n\nThis fix implement the necessary weights changes, aligning to Jetstream\nPytorch, and tests expected output has been modified accordingly.","shortMessageHtmlLink":"fix(Jetstream Pt): correct generation"}},{"before":"a8b195c8f78c002ff5ad0885d99176ea94be525c","after":"7b9c0ca6610ce68bd470e230456c3d1ed435e5b3","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-18T10:50:17.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"fix(Jetstream Pt): correct generation\n\nText generation was not correct because the weights in the model were\nnot correctly loaded. This is not something that it was easy to spot\njust looking at few tokens generated, and it was something that it was\nactually fixed already in the Jetstream/Pytorch code, but the fix hadn't\nbeen ported to optimum-tpu.\n\nThis fix implement the necessary weights changes, aligning to Jetstream\nPytorch.","shortMessageHtmlLink":"fix(Jetstream Pt): correct generation"}},{"before":null,"after":"a8b195c8f78c002ff5ad0885d99176ea94be525c","ref":"refs/heads/investigate-generation","pushedAt":"2024-09-17T15:45:53.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP fix(Jetstream Pt): correct model weight loading","shortMessageHtmlLink":"WIP fix(Jetstream Pt): correct model weight loading"}},{"before":"c7e0bcb6f235150eca055b97d66eefeb5037523d","after":null,"ref":"refs/heads/fix-extra-token","pushedAt":"2024-09-17T12:17:08.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"}},{"before":"03b657362139f63f295b6bd8ecd68c68cf5fadb3","after":"4265e1326c2e0d269e68d61a971c723d3b318ab1","ref":"refs/heads/main","pushedAt":"2024-09-17T12:17:04.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"Correct extra token, start preparing docker image for TGI/Jetstream Pt (#93)\n\n* fix(Jetstream Pt): remove extra token in decode\r\n\r\nJetstream's `generate` function returns input token as result token. The\r\nnext token is instead available in the decode_state, so this change uses\r\nthis instead.\r\n\r\n* fix(engine): set batch_size and sequence_length\r\n\r\n* fix(Jetstream PT): correct warmup internal params\r\n\r\n* test(tgi): added a warmup test\r\n\r\n* chore(jetstream pt): check input type in decode\r\n\r\n* fix(token selector): seed can be a very large number\r\n\r\nBefore, we could have an error is the seed was bigger than a 64 bit\r\nnumber.\r\n\r\n* fix(Jetstream PT): handle slot's seed in a clean way\r\n\r\n* feat(docker): TGI image now include Jetstream Pytorch dependencies\r\n\r\nThis allows testing TGI images with Jetstream Pytorch.\r\n\r\n* fix(Jetstream Pt): batch returned in prefill initialized to None\r\n\r\nThis is required when there are no more tokens generated after prefill.\r\n\r\n* feat(Jetstream Pt): speed-up prefill by avoiding redundant compilation\r\n\r\nA new slot is created at each prefill request, and its selector is\r\npassed as argument to a jitted function. The problem is that each new\r\nslot has a new signature, even if the contents are the same. The\r\nsolution is to wrap that in a singleton slot object for the prefill, so\r\nthe compiler will always see the same object and stop recompiling.\r\n\r\n* chore(generator): use prefill bucket sizes defined in Jetstream","shortMessageHtmlLink":"Correct extra token, start preparing docker image for TGI/Jetstream Pt ("}},{"before":"5e66bf7e7818f7d585cd6db8de1768c2b94cfe22","after":"c7e0bcb6f235150eca055b97d66eefeb5037523d","ref":"refs/heads/fix-extra-token","pushedAt":"2024-09-17T08:18:43.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"chore(generator): use prefill bucket sizes defined in Jetstream","shortMessageHtmlLink":"chore(generator): use prefill bucket sizes defined in Jetstream"}},{"before":"32095e4ec7a1f444d678bab26089f98ae1a02050","after":"5e66bf7e7818f7d585cd6db8de1768c2b94cfe22","ref":"refs/heads/fix-extra-token","pushedAt":"2024-09-16T14:56:29.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"feat(Jetstream Pt): speed-up prefill by avoiding redundant compilation\n\nA new slot is created at each prefill request, and its selector is\npassed as argument to a jitted function. The problem is that each new\nslot has a new signature, even if the contents are the same. The\nsolution is to wrap that in a singleton slot object for the prefill, so\nthe compiler will always see the same object and stop recompiling.","shortMessageHtmlLink":"feat(Jetstream Pt): speed-up prefill by avoiding redundant compilation"}},{"before":"75398118320281c4d8b3f2f7f982e40a4691bde0","after":"32095e4ec7a1f444d678bab26089f98ae1a02050","ref":"refs/heads/fix-extra-token","pushedAt":"2024-09-13T14:54:27.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"feat(Jetstream Pt): speed-up prefill by avoiding redundant compilation\n\nA new slot is created at each prefill request, and its selector is\npassed as argument to a jitted function. The problem is that each new\nslot has a new signature, even if the contents are the same. The\nsolution is to wrap that in a singleton slot object for the prefill, so\nthe compiler will always see the same object and stop recompiling.","shortMessageHtmlLink":"feat(Jetstream Pt): speed-up prefill by avoiding redundant compilation"}},{"before":"a84803fd6c21a223333ce684aae56984a9a21361","after":"75398118320281c4d8b3f2f7f982e40a4691bde0","ref":"refs/heads/fix-extra-token","pushedAt":"2024-09-13T14:20:06.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"feat(Jetstream Pt): speed-up prefill by avoiding redundant compilation\n\nA new slot is created at each prefill request, and its selector is\npassed as argument to a jitted function. The problem is that each new\nslot has a new signature, even if the contents are the same. The\nsolution is to wrap that in a singleton slot object for the prefill, so\nthe compiler will always see the same object and stop recompiling.","shortMessageHtmlLink":"feat(Jetstream Pt): speed-up prefill by avoiding redundant compilation"}},{"before":"ef41b2cd3bae323272c522c697530d331f5eb106","after":"a84803fd6c21a223333ce684aae56984a9a21361","ref":"refs/heads/fix-extra-token","pushedAt":"2024-09-13T13:48:13.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"fix(Jetstream Pt): batch returned in prefill initialized to None\n\nThis is required when there are no more tokens generated after prefill.","shortMessageHtmlLink":"fix(Jetstream Pt): batch returned in prefill initialized to None"}},{"before":"1b3eb1840cdbffe78f3daae02120188cbcb7de97","after":"ef41b2cd3bae323272c522c697530d331f5eb106","ref":"refs/heads/fix-extra-token","pushedAt":"2024-09-12T15:38:45.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP","shortMessageHtmlLink":"WIP"}},{"before":"ed796f95e8f12bc5ca0eac84113cadd129cfd237","after":null,"ref":"refs/heads/update-jetstream-pt","pushedAt":"2024-09-12T15:22:45.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"}},{"before":"b25e973a8f18d93d7a2080b2d46b384adf433e26","after":"03b657362139f63f295b6bd8ecd68c68cf5fadb3","ref":"refs/heads/main","pushedAt":"2024-09-12T15:22:36.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"☝️ Update Jetstream Pytorch revision (#91)\n\n* feat(Jetstream PT): update git version\r\n\r\n* feat(jetstream_pt): drop prefill_ex method\r\n\r\n* feat(jetstream_pt): drop generate_ex method\r\n\r\nNow that decode takes a specialized method this is not necessary\r\nanymore.\r\n\r\n* refactor(Jetstream PT): drop HfEngine\r\n\r\nNow that sampling can be passed as parameter to prefill and generate,\r\nthe custom engine is not required anymore.\r\n\r\n* review(Jetstream Pt): avoid double iteration on slots","shortMessageHtmlLink":"☝️ Update Jetstream Pytorch revision (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2516649887\" data-permission-text=\"Title is private\" data-url=\"https://github.com/huggingface/optimum-tpu/issues/91\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/huggingface/optimum-tpu/pull/91/hovercard\" href=\"https://github.com/huggingface/optimum-tpu/pull/91\">#91</a>)"}},{"before":"4674367d17bd4ec7296e6ab73a093a5936b12198","after":"ed796f95e8f12bc5ca0eac84113cadd129cfd237","ref":"refs/heads/update-jetstream-pt","pushedAt":"2024-09-12T15:08:04.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"review(Jetstream Pt): avoid double iteration on slots","shortMessageHtmlLink":"review(Jetstream Pt): avoid double iteration on slots"}},{"before":"6f00b82d6eb0829a5741645fd08b9b12b61db85f","after":"694db114cfc755f6fdf4aa674140711b87517890","ref":"refs/heads/quick-ci-test","pushedAt":"2024-09-11T11:47:34.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP","shortMessageHtmlLink":"WIP"}},{"before":"461e2b1b5274465eb75b7571daca3c70f16b8e20","after":"6f00b82d6eb0829a5741645fd08b9b12b61db85f","ref":"refs/heads/quick-ci-test","pushedAt":"2024-09-11T11:39:38.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP","shortMessageHtmlLink":"WIP"}},{"before":"1abc373ef90474ebccde6ed70c1f59c2d41695bc","after":"461e2b1b5274465eb75b7571daca3c70f16b8e20","ref":"refs/heads/quick-ci-test","pushedAt":"2024-09-11T11:38:08.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP","shortMessageHtmlLink":"WIP"}},{"before":"10f857c4eadab57ad24a2d0a7c862c8aac5de4f5","after":"1abc373ef90474ebccde6ed70c1f59c2d41695bc","ref":"refs/heads/quick-ci-test","pushedAt":"2024-09-11T11:36:04.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP","shortMessageHtmlLink":"WIP"}},{"before":"51ec0e38f55cd13539ddf4b4cefdc33d3c5f3372","after":"10f857c4eadab57ad24a2d0a7c862c8aac5de4f5","ref":"refs/heads/quick-ci-test","pushedAt":"2024-09-11T11:35:39.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"Merge pull request #2 from Bihan/add_llama_31_support_backup\n\nAdd llama 31 support updated","shortMessageHtmlLink":"Merge pull request <a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2156787303\" data-permission-text=\"Title is private\" data-url=\"https://github.com/huggingface/optimum-tpu/issues/2\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/huggingface/optimum-tpu/pull/2/hovercard\" href=\"https://github.com/huggingface/optimum-tpu/pull/2\">#2</a> from Bihan/add_llama_31_support_backup"}},{"before":"a66154a2cdf60e5a90503a14b98fa0d18a22678d","after":"51ec0e38f55cd13539ddf4b4cefdc33d3c5f3372","ref":"refs/heads/quick-ci-test","pushedAt":"2024-09-11T11:30:52.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP test ci","shortMessageHtmlLink":"WIP test ci"}},{"before":"ff1b37130c4f40d50cd22d53d9ea3a41a11b57aa","after":"a66154a2cdf60e5a90503a14b98fa0d18a22678d","ref":"refs/heads/quick-ci-test","pushedAt":"2024-09-11T11:22:00.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"tengomucho","name":"Alvaro Moran","path":"/tengomucho","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6949769?s=80&v=4"},"commit":{"message":"WIP","shortMessageHtmlLink":"WIP"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEuw2fAgA","startCursor":null,"endCursor":null}},"title":"Activity · huggingface/optimum-tpu"}