Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[v1.2.0] Release Tracker #23555

Closed
soumith opened this issue Jul 30, 2019 · 33 comments
Closed

[v1.2.0] Release Tracker #23555

soumith opened this issue Jul 30, 2019 · 33 comments
Labels
triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Comments

@soumith
Copy link
Member

soumith commented Jul 30, 2019

The branch v1.2.0 has been cut.

If you need any particular patches onto this branch, please comment below and send a PR to v1.2.0 branch (instead of master).


Current PRs open against the v1.2.0 branch: https://github.com/pytorch/pytorch/pulls?utf8=%E2%9C%93&q=is%3Apr+is%3Aopen+base%3Av1.2.0

@ezyang
Copy link
Contributor

ezyang commented Jul 30, 2019

@brianjo Can you make sure #23553 gets into 1.2.0 after you land it on master

#23376 should get to 1.2.0 too.

Both of these doc PRs are not release blocking, because we can repush the docs when necessary.


Status: #235353 merged into 1.2.0 via 214dc02 and #23376 merged via 3e8e88b

@jerryzh168
Copy link
Contributor

jerryzh168 commented Jul 30, 2019

this pr need to be landed to 1.2: #23452 (landed in master)
cherry-picked: #23604


Status: Merged at c3bad0d

@VitalyFedyunin VitalyFedyunin added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Jul 30, 2019
@sidazhang
Copy link

Is this going to make it into 1.2? #23241

@soumith
Copy link
Member Author

soumith commented Jul 31, 2019

@sidazhang the pickler API is explicitly not making into 1.2, because @suo thinks it needs to bake for a few days in nightlies to test stability

@vishwakftw
Copy link
Contributor

vishwakftw commented Jul 31, 2019

#23591 could also be included in 1.2 (and in master)


Status: Merged at d6df957

@zou3519
Copy link
Contributor

zou3519 commented Jul 31, 2019

#23596 makes some changes to the doc push so we will be able to set up the docs job for v1.2.0 later.


Status: Merged at 49e32ff

@ssnl
Copy link
Collaborator

ssnl commented Jul 31, 2019

If #23587 will be accepted, I would love to see it in 1.2 too.


Status: original PR still under review

@soumith
Copy link
Member Author

soumith commented Jul 31, 2019

#23298 needs to be in 1.2 as well.


Status: Merged at 6f5fb78

@vishwakftw
Copy link
Contributor

vishwakftw commented Jul 31, 2019

Could #22909 be accepted in v1.2.0 as well?


Status: Merged at f0a65e6

@suo
Copy link
Member

suo commented Jul 31, 2019

#23630 should go in 1.2 as well


Status: Merged at db37022

@soumith soumith pinned this issue Aug 1, 2019
@soumith
Copy link
Member Author

soumith commented Aug 1, 2019

#23629 should be cherry-picked as well


Status: Merged at ccc7b5c

@ssnl
Copy link
Collaborator

ssnl commented Aug 1, 2019

#23646 needs to be cherry-picked too


Status: Merged at bb4ff00

@suo
Copy link
Member

suo commented Aug 1, 2019

#23681 will need to be cherry-picked (fixes a BC breakage with older JIT models)


Status: Merged at a0e9dd3

@driazati
Copy link
Contributor

driazati commented Aug 1, 2019

#23686 is open against v1.2.0 and should be merged


Status: Merged at 4f52116

@wanchaol
Copy link
Contributor

wanchaol commented Aug 2, 2019

#23700 is also opened against v1.2.0 and needs to be merged (More complete support on JIT GRU and LSTM with PackedSequence)


Status: Merged at 5e36bec

@ssnl
Copy link
Collaborator

ssnl commented Aug 2, 2019

#23709


Status: Merged at 7e327f2

@vishwakftw
Copy link
Contributor

vishwakftw commented Aug 2, 2019

#23740 as well


Status: Merged at 47828fd

@xuhdev
Copy link
Collaborator

xuhdev commented Aug 2, 2019

#23528 is in v1.2.0 but it causes a build regression. It is currently reverted in master. We should either revert it or have its fix (#23568) merged in v1.2.0


Status: Merged at 4d1d843

@ssnl
Copy link
Collaborator

ssnl commented Aug 3, 2019

#23762 too


Status: Merged at 1ac19ea

@izdeby
Copy link
Contributor

izdeby commented Aug 5, 2019

#23789


Status: Merged at 3314b57

@izdeby
Copy link
Contributor

izdeby commented Aug 5, 2019

#23811


Status: PR merged at 060f677

@raghuramank100 raghuramank100 unpinned this issue Aug 5, 2019
@vishwakftw
Copy link
Contributor

vishwakftw commented Aug 6, 2019

#23855 is a doc string fix for argmax.


Status: Merged at 0d7d1d0

@gchanan
Copy link
Contributor

gchanan commented Aug 6, 2019

#23951 is a docstring fix for torch.frac.


Status: PR merged at 2ad6f42

@gchanan
Copy link
Contributor

gchanan commented Aug 7, 2019

#23952 fixes an optimizer-step order for CosineAnnealingWarmRestarts.


Status: PR merged at cdb032e

@gchanan
Copy link
Contributor

gchanan commented Aug 7, 2019

#23955 for fixing some bool bugs.


Status: PR merged at 805556c

@gchanan gchanan pinned this issue Aug 7, 2019
@ezyang
Copy link
Contributor

ezyang commented Aug 7, 2019

#23954 for fixing C++ API for pooling operations when passed size-1 stride (this will be a stack)


Status: PR merged at 7d13820

@vishwakftw
Copy link
Contributor

vishwakftw commented Aug 8, 2019

#23997 fixes a performance regression in triangular_solve when number of batches is 1.


Status: PR merged at 11d7e8d

@barrh
Copy link
Contributor

barrh commented Aug 8, 2019

@gchanan Would it be possible to upload the latest v1.2 release for CUDA 10 by the explicit name 1.2.0+cu10 , to pytorch.org/whl/ ?
currently, I see: 1.2.0, 1.2.0+cpu, 1.2.0+cu92

thanks

@ezyang
Copy link
Contributor

ezyang commented Aug 9, 2019

Hi @barrh, it's possible, but could you explain in more detail why you want this?

@Nikronic
Copy link
Contributor

Nikronic commented Aug 9, 2019

Hi, v1.2.0 is not available for Windows clients on pip.
https://discuss.pytorch.org/t/pytorch-1-2-windows/52959

@ezyang
Copy link
Contributor

ezyang commented Aug 9, 2019

@Nikronic Yeah, there was some trouble with the builds. Hopefully they'll be up today.

@peterjc123
Copy link
Collaborator

@Nikronic #24141

@barrh
Copy link
Contributor

barrh commented Aug 11, 2019

@ezyang It will be easier to identify which devices are supported.
Assuming this naming scheme will stay with us on future releases, giving full names will help thwart any ambiguity.

pytorch1.1 -> default cuda is v9
pytorch1.2 -> default cuda is v10
pytorch1.3 -> v10.1 ? v11?

@soumith soumith unpinned this issue Aug 11, 2019
@soumith soumith closed this as completed Sep 28, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Projects
None yet
Development

No branches or pull requests