Skip to content

Fix quantized references #8073

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.

Sign up for GitHub

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

Jump to bottom
Merged
merged 2 commits into from
Oct 27, 2023
Merged

Fix quantized references #8073

merged 2 commits into from
Oct 27, 2023

Conversation

@NicolasHug
Copy link
Member

@NicolasHug NicolasHug commented Oct 27, 2023

Fixes #8069
@pytorch-bot
Copy link

pytorch-bot bot commented Oct 27, 2023 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/8073

Note: Links to docs will display an error until the docs builds have been completed.

❌ 12 New Failures, 15 Unrelated Failures

As of commit 752cf09 with merge base a63f98a (image):

NEW FAILURES - The following jobs have failed:

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

This comment was automatically generated by Dr. CI and updates every 15 minutes.
NicolasHug
NicolasHug commented Oct 27, 2023
View reviewed changes
references/classification/train_quantization.py
Comment on lines +260 to +261
parser.add_argument("--backend", default="PIL", type=str.lower, help="PIL or tensor - case insensitive")
parser.add_argument("--use-v2", action="store_true", help="Use V2 transforms")
Copy link
Member Author

@NicolasHug NicolasHug Oct 27, 2023

Choose a reason for hiding this comment

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

This is the actual fix. These fields are accessed in

vision/references/classification/train.py

Lines 147 to 148 in a63f98a

backend=args.backend,
use_v2=args.use_v2,

but since the args parameter comes from this script (and not from train.py), they need to be defined here as well.

Since the --backend parameter already existed in this file, I decided to break BC and rename it into qbackend.
Copy link
Contributor

@pmeier pmeier Oct 27, 2023

Choose a reason for hiding this comment

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

This needs to be reflected in the README, e.g.

vision/references/classification/README.md

Line 292 in a63f98a

5. backend: 'fbgemm'

@NicolasHug NicolasHug requested review from pmeier and vfdev-5 October 27, 2023 11:16
pmeier
pmeier approved these changes Oct 27, 2023
View reviewed changes
Copy link
Contributor

@pmeier pmeier left a comment

Choose a reason for hiding this comment

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

Thanks!
references/classification/train_quantization.py
Comment on lines +260 to +261
parser.add_argument("--backend", default="PIL", type=str.lower, help="PIL or tensor - case insensitive")
parser.add_argument("--use-v2", action="store_true", help="Use V2 transforms")
Copy link
Contributor

@pmeier pmeier Oct 27, 2023

Choose a reason for hiding this comment

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

This needs to be reflected in the README, e.g.

vision/references/classification/README.md

Line 292 in a63f98a

5. backend: 'fbgemm'

@NicolasHug NicolasHug merged commit 96d2ce9 into pytorch:main Oct 27, 2023
@NicolasHug NicolasHug deleted the aljnalefjnajnaef branch October 27, 2023 12:26
facebook-github-bot pushed a commit that referenced this pull request Nov 13, 2023
@NicolasHug @facebook-github-bot
[fbsync] Fix quantized references (#8073)
8b381f3 
Reviewed By: vmoens Differential Revision: D50789093 fbshipit-source-id: 17b6840f89063eeef1fb429ff04817f46e5919eb
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug cla signed module: reference scripts

3 participants

@NicolasHug @pmeier @facebook-github-bot