symintify rand and randint functions and meta suport for randint #86358
Conversation
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/86358
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 FailuresAs of commit 561a01e: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
github-actions
bot
requested review from
albanD,
antoniojkim,
bdhirsh,
Chillee,
ezyang,
Krovatkin,
miladm and
wconstab
| c10::impl::ExcludeDispatchKeyGuard guard(DispatchKey::FuncTorchVmapMode); | ||
| auto maybe_layer = maybeCurrentDynamicLayer(); | ||
| VmapDimVector shapeVec(1, maybe_layer->batchSize()); | ||
| c10::SmallVector<SymInt> shapeVec(1, maybe_layer->batchSize()); |
There was a problem hiding this comment.
Not sure if this is ok given @ezyang 's latest ambitions?
There was a problem hiding this comment.
Use SymDimVector instead, but this is fine, I gave up on changing SymIntArrayRef
torch/_meta_registrations.py
Outdated
| @@ -80,6 +80,9 @@ def meta_randperm(n, *, generator=None, out): | |||
| assert out.ndim == 1 and out.size(0) == n | |||
| return out | |||
|
|
|||
| @register_meta(aten.randint.default) | |||
| def meta_randperm(high, size, dtype, layout, device, pin_memory): | |||
| return torch.empty(size, dtype=dtype, device=device) | |||
There was a problem hiding this comment.
Why not pass layout and pin memory?
There was a problem hiding this comment.
oh yeah that needs to be updated
There was a problem hiding this comment.
one idiom for factory functions is to just pass along the entire **kwargs
…andint" [ghstack-poisoned]
…andint" [ghstack-poisoned]
|
@pytorchbot merge -g |
Merge startedYour change will be merged once all checks on your PR pass since you used the green (-g) flag (ETA: 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Merge failedReason: The following mandatory check(s) failed (Rule Dig deeper by viewing the failures on hud Details for Dev Infra teamRaised by workflow job |
…andint" [ghstack-poisoned]
…andint" [ghstack-poisoned]
…andint" [ghstack-poisoned]
…andint" [ghstack-poisoned]
…andint" [ghstack-poisoned]
|
@pytorchbot merge -g |
Merge startedYour change will be merged once all checks on your PR pass since you used the green (-g) flag (ETA: 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Merge failedReason: 2 additional jobs have failed, first few of them are: trunk ,trunk / linux-focal-rocm5.2-py3.7 / test (default, 2, 2, linux.rocm.gpu) Details for Dev Infra teamRaised by workflow job |
|
@pytorchbot merge -f "unrelated rocm failure" |
Merge startedYour change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
|
Hey @anjali411. |
ghstack-source-id: 559ca23 Pull Request resolved: pytorch/pytorch#86358
Stack from ghstack (oldest at bottom):