Skip to content
This repository has been archived by the owner on Sep 18, 2024. It is now read-only.

(docs) Release note (v2.10) #5221

Merged
merged 15 commits into from
Nov 14, 2022
Merged

(docs) Release note (v2.10) #5221

merged 15 commits into from
Nov 14, 2022

Conversation

Louis-J
Copy link
Contributor

@Louis-J Louis-J commented Nov 11, 2022

Description

Test Options

  • fast test
  • full test - HPO
  • full test - NAS
  • full test - compression

Checklist

  • test case
  • doc

How to test

@Louis-J Louis-J changed the base branch from master to release/v2.10 November 11, 2022 08:53
""""""""""""

* Modelspeedup does not support non-tensor intermediate variables
* The training service runs much slower after a while.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Don't include this line. There's nothing to do with NNI.

Neural Architecture Search
^^^^^^^^^^^^^^^^^^^^^^^^^^

* Added deduplication for evolutionary search.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

deduplication -> trial deduplication

^^^^^^^^^^^^^^^^^^^^^^^^^^

* Added deduplication for evolutionary search.
* Fixed the racing issue in RL strategy submit model.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

submit model -> on submitting models?

* Added deduplication for evolutionary search.
* Fixed the racing issue in RL strategy submit model.
* Fixed an issue introduced by the trial recovery feature.
* Fixed import error of ``Lightning`` in nas.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Lightning should be PyTorch Lightning. nas should be NAS

Compression
"""""""""""

* #3514
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

a little strange to simply put an id here

* #3514
* Fixed the bug that speedup ``rand_like_with_shape`` is easy to overflow when ``dtype=torch.int8``.
* Fixed the bug when calling ``torch._C.parse_schema`` in pytorch < 1.9.0.
* Fixed the propagation error with view tensors in speedup.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

-> propagation error of torch.view operator ?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

view tensors can be produced by many operators(30+, such as expand, t, or detach).

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Got it. Is "...error with viewing (e.g., torch.view, torch.reshape) tensors..." better?

Hyper-parameter optimization
""""""""""""""""""""""""""""

* Fixed a dependency issue of Anneal tuner by change Anneal tuner dependency to optional.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

change -> changing

""""""""""""""""""""""""""""

* Fixed a dependency issue of Anneal tuner by change Anneal tuner dependency to optional.
* Fixed a bug that tuner might loss connection in long experiments.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

loss -> lose

Web portal
""""""""""

* Fixed an error in HPO experiment hyper-parameters page.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

what is this error? better to very briefly describe it


Hyper-parameter optimization
""""""""""""""""""""""""""""

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

could add one more: Supported rerunning (resuming) the interrupted trials induced by the termination of an NNI experiment when resuming this experiment.

@Louis-J Louis-J marked this pull request as ready for review November 14, 2022 03:53
docs/source/release.rst Outdated Show resolved Hide resolved
@Louis-J Louis-J closed this Nov 14, 2022
@Louis-J Louis-J reopened this Nov 14, 2022
@Louis-J Louis-J closed this Nov 14, 2022
@Louis-J Louis-J reopened this Nov 14, 2022
@Louis-J Louis-J merged commit c31d257 into microsoft:v2.10 Nov 14, 2022
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants