Mikhail Druzhinin
@Dipetm
Followers
47
Following
65
Media
1
Statuses
16
https://t.co/MNPclDHVb1 PIL to @albumentations switch doubled @LightlyAI training speed and pushed GPU utilization to 100%
lightly.ai
Switching from Pillow to Albumentations and Torchvision doubled our data preprocessing speed, pushing GPU utilization to nearly 100%. Read how we optimized our pipeline for faster training.
0
2
4
Release 1.4.5 - Now can clip bounding boxes if the extend outside of the image - Added SelectiveChannelTransform that allows applying augmentations to a subset of channels. Useful when working with multispectral images. + a few bugfixes https://t.co/9eqr1Ni8Bb
github.com
Support our work Highlights Deprecations Improvements and bug fixes Support Our Work Love the library? You can contribute to its development by becoming a sponsor for the library. Your support i...
0
3
9
Added @kaggle type gamification to #Python #opensource development. Leaderboard. Top 1.5% (8000 out of 520,000) packages Metric - the number of downloads from @pypi in the past 30 days. https://t.co/E0LdXF8TWz
0
1
7
Version 1.4.2 of Albumentations is available. Release Highlights: - New transform - ChromeAbberation. - The MixUp transform can now return the mixing coefficient - Fixed an issue with PadIfNeeded serialization. - Documentation improvements. #imageprocessing #deeplearning
2
2
3
In version 1.4.2, we added the ChromaticAberration transform. Could be helpful in: - High-contrast scenes - Wide apertures photography - Telephoto lens usage - Digital zooming - Underwater and action photography https://t.co/LMJOGqp723
dev.to
Albumentations 1.4.2 adds the Chromatic Aberration transform. This feature simulates the common lens...
0
2
2
Release 1.4.1 https://t.co/tUEKTCA2ps The main highlight - the MixUp transform, which could be applied to - image - mask - global_label Notebook with the example of the application: https://t.co/JcBUvV9036
0
2
7
Big thanks @roboflow for sponsoring #albumentations at @github
#opensource needs funding to move forward and any financial contribution provides a lot of value. https://t.co/ylWKOAEcfr
github.com
Albumentations is a fast and flexible library for image augmentation, designed to support training of computer vision models through a broad set of configurable transformation techniques.
1
1
10
#Albumentations 1.3.0 released. Five new transforms and numerous improvements, speedups, and bug fixes. Added bonding box support to the GridDistortion, ElasticTransform, and OpticalDistortion. More info at https://t.co/Ca14TUJ8fX
0
8
33
New release of the best image augmentations library Albumentations - 1.2.0 - Four new transforms - Improved key points support - Improved documentation - Various bug fixes Stats: - 280k downloads per month - 10.4k stars at GitHub. More details: https://t.co/sMQA92YkBp
0
3
13
#Albumentations 1.1.0 is out! The new release of a fast and flexible library for image augmentation includes new transformations and improvements for the current ones. Thanks to @Dipetm @viglovikov @cvtalks @AlBuslaev and all our contributors. https://t.co/NMwgwx6NpT
1
6
12
The new version of our library for fast image generation of people who do not exist was released. x2 Better quality of generated images is here (= python library: https://t.co/iGJpa4pO7T training tool: https://t.co/bwRH2jSZtZ demo video: https://t.co/jXKLLsNd01
0
36
110
Version 1.0.0 of the most advanced template for Python projects on @github has been released! New generation parameters, updated commands in Makefile, dependencies refactoring. All this happened thanks to feedback from users. Check it now 📈
github.com
🚀 Your next Python package needs a bleeding-edge project structure. - TezRomacH/python-package-template
I released my cookiecutter template for your Python package with a bleeding-edge project structure. - @github Actions: CI and Bots - pre-commit with black, isort, pyupgrade - codestyle/security with mypy, pytest, darglint, etc - many more, check the link https://t.co/9cYaeSfgP9
0
12
30
#Albumentations 1.0.0 has been released! New version contains: - 10 new transforms - bug fixes etc See the release notes for details https://t.co/QLEfhmpYgk Thanks a lot to core team @viglovikov @creaf @cvtalks @AlBuslaev and everyone who helps improve the library
1
23
77
I am thrilled that image Augmentation library #albumentations that was born out of @kaggle competitions by @AlBuslaev @creaf @cvtalks @Dipetm and @viglovikov became a part of the @PyTorch ecosystem! https://t.co/oGokF8ZxeG
landscape.pytorch.org
4
97
413
@AlBuslaev @creaf @cvtalks @dipetm and @viglovikov are happy to announce [#albumentations 0.4.0] 10 new transforms 7 extra transforms got keypoints support 3+ channel images and masks debug mode YOLO bounding boxes speedups etc Read more here: https://t.co/1ghLhMxqdR
2
12
39