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

[core] move transformers to setup_utils, bump dependency version #3984

Merged
merged 4 commits into from
Mar 21, 2024

Conversation

canerturkmen
Copy link
Contributor

Issue #, if available:

Description of changes:

Chronos implementation (cf #3978) requires transformers>=4.36. This PR moves transformers and accelerate version ranges to _setup_utils.py and bumps the version.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

@canerturkmen canerturkmen added module: timeseries related to the timeseries module module: multimodal run-multi-gpu Run multimodal multi-gpu tests labels Mar 15, 2024
@canerturkmen canerturkmen added the priority: 0 Maximum priority label Mar 15, 2024
@canerturkmen canerturkmen added this to the 1.1 Release milestone Mar 15, 2024
@canerturkmen
Copy link
Contributor Author

In Chronos, we would like to depend on optimum-intel for fast CPU inference, and this requires transformers>=4.36. We can ship without optimum-intel support too, but this results in 2-3x slower inference on some models.

Currently, bumping the transformers version appears to be breaking a specific unit test. If this is easy to mitigate for MM, then we can include that fix here or in another PR to enable Chronos.

Looking forward to your comments.

@tonyhoo
Copy link
Collaborator

tonyhoo commented Mar 15, 2024

@prateekdesai04 we should test this change with torch update

@@ -55,7 +57,9 @@
],
}

extras_require["all"] = []
extras_require["all"] = [
"optimum[openvino,onnxruntime,nncf]>=1.17,<1.18", # for faster CPU inference in pretrained models
Copy link
Collaborator

Choose a reason for hiding this comment

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

Do we need neuron suppport to speed up inference on Inf2 architecture?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Not at the moment, though this will come in v1.2.

@prateekdesai04
Copy link
Contributor

@prateekdesai04 we should test this change with torch update

@tonyhoo Just pushed the latest Torch + CUDA image, running this PR #3982
Once successful we can merge it and then run Caner's PR with the latest changes

Copy link

Job PR-3984-a4b7229 is done.
Docs are uploaded to http://autogluon-staging.s3-website-us-west-2.amazonaws.com/PR-3984/a4b7229/index.html

@canerturkmen canerturkmen marked this pull request as draft March 16, 2024 12:08
@canerturkmen
Copy link
Contributor Author

canerturkmen commented Mar 16, 2024

Converted to draft, pending the PyTorch upgrade. and thanks @zhiqiangdon !

@canerturkmen canerturkmen marked this pull request as ready for review March 19, 2024 06:12
@yinweisu
Copy link
Collaborator

Previous CI Run Current CI Run

Copy link

Job PR-3984-a0635c5 is done.
Docs are uploaded to http://autogluon-staging.s3-website-us-west-2.amazonaws.com/PR-3984/a0635c5/index.html

@yinweisu
Copy link
Collaborator

Previous CI Run Current CI Run

@canerturkmen
Copy link
Contributor Author

@zhiqiangdon for your review of test fails

@yinweisu
Copy link
Collaborator

Previous CI Run Current CI Run

@yinweisu
Copy link
Collaborator

Previous CI Run Current CI Run

Copy link

Job PR-3984-314e8c3 is done.
Docs are uploaded to http://autogluon-staging.s3-website-us-west-2.amazonaws.com/PR-3984/314e8c3/index.html

Copy link
Contributor

@zhiqiangdon zhiqiangdon left a comment

Choose a reason for hiding this comment

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

LGTM

@canerturkmen canerturkmen merged commit 4d5a054 into autogluon:master Mar 21, 2024
35 checks passed
@canerturkmen canerturkmen deleted the bump-transformers-version branch March 21, 2024 05:58
ddelange added a commit to ddelange/autogluon that referenced this pull request Mar 21, 2024
…tch-4

* 'master' of https://github.com/awslabs/autogluon: (46 commits)
  [core] move transformers to setup_utils, bump dependency version (autogluon#3984)
  [AutoMM] Fix one lightning upgrade issue (autogluon#3991)
  [CI][Feature] Create a package version table (autogluon#3972)
  [v.1.1][Upgrade] PyTorch 2.1 and CUDA 12.1 upgrade (autogluon#3982)
  [WIP] Code implementation of Conv-LoRA (autogluon#3933)
  [timeseries] Ensure that all metrics handle missing values in the target (autogluon#3966)
  [timeseries] Fix path and device bugs (autogluon#3979)
  [AutoMM]Remove grounding-dino (autogluon#3974)
  [Docs] Update install modules content (autogluon#3976)
  Add note on pd.to_datetime (autogluon#3975)
  [AutoMM] Improve DINO performance (autogluon#3970)
  Minor correction in differ to pick correct environment (autogluon#3968)
  Fix windows python 3.11 issue by removing ray (autogluon#3956)
  [CI][Feature] Package Version Comparator (autogluon#3962)
  [timeseries] Add support for categorical covariates (autogluon#3874)
  [timeseries] Add method for plotting forecasts (autogluon#3889)
  Update conf.py copyright to reflect current year (autogluon#3932)
  [Timeseries][CI]Refactor CI to skip AutoMM and Tabular tests w.r.t timeseries changes (autogluon#3942)
  Fix HPO crash in memory check (autogluon#3931)
  [AutoMM][CI] Capping scikit-learn to avoid HPO test failure (autogluon#3947)
  ...
prateekdesai04 pushed a commit to prateekdesai04/autogluon that referenced this pull request Apr 3, 2024
…ogluon#3984)

Co-authored-by: Zhiqiang Tang <zhiqiang.tang@rutgers.edu>
LennartPurucker pushed a commit to LennartPurucker/autogluon that referenced this pull request Jun 1, 2024
…ogluon#3984)

Co-authored-by: Zhiqiang Tang <zhiqiang.tang@rutgers.edu>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
module: multimodal module: timeseries related to the timeseries module priority: 0 Maximum priority run-multi-gpu Run multimodal multi-gpu tests
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

5 participants