Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/7670
Note: Links to docs will display an error until the docs builds have been completed. ⏳ No Failures, 68 PendingAs of commit 480356f with merge base 01d526f ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
jackzhxng
left a comment
There was a problem hiding this comment.
Why is it not in the bucket anymore - do we evict old versions? Ideally we only update the TorchTune pin out of necessity since the test coverage for this model is not completely there yet
|
@dvorjackz - yeah i think we only keep like 90 days (or 60 days) in our nightlies https://download.pytorch.org/whl/nightly/torchtune/ perhaps to save on storage spaces and for maintenance. |
|
@pytorchbot cherry-pick --onto release/0.5 -c release |
(cherry picked from commit a18f6e8)
Cherry picking #7670The cherry pick PR is at #7710 The following tracker issues are updated: Details for Dev Infra teamRaised by workflow job |
(cherry picked from commit a18f6e8)
(cherry picked from commit a18f6e8)
* Install torchao by default (#7470) Summary: Now that torchao is not experimental anymore, we can rely on torchao directly via git submodule. Part of #7460 (cherry picked from commit a861294) * Upgrade numpy version requirement to 2.0 (#7599) Upgrade numpy version requirement to Summary: 1.21 is too old now. For 3.10 or above, we should be able to use 2.0.0 We are already on CoreMLTools 8.1, which is https://github.com/apple/coremltools/blob/8.1/reqs/build.pip We had to upgrade to the latest transformers as well (cherry picked from commit a7b5297) * Update torchtune pin (#7670) (cherry picked from commit a18f6e8) * [Arm] TOSA reference model update * Install release version of torchtune --------- Co-authored-by: Sicheng Stephen Jia <ssjia@meta.com>
Summary
Current torchtune pin is not in the S3 bucket anymore. https://download.pytorch.org/whl/nightly/torchtune/
Updating to latest
Test
CI