ternaus

8 ms·

1.
by ternaus 2mo ago
Audiobooks is the magic trick. I listen them in the gym between sets, I listen them in the car, in the plane, in taxi, doing chores at home and in any other place where you cannot do more useful things, and you cannot read as well. Five min
2.
by ternaus 3mo ago
Same here, 15k stars, 150M downloads and never heard back.
3.
by ternaus 3mo ago
To the question about OpenCV peformance to measure JPEG image decoding sequentially and as a part of the PyTorch Dataloader. TL;DR OpenCV is fast, but torchvision is faster. https://arxiv.org/abs/2605.08731
4.
by ternaus 3mo ago
Image augmentations library Albumentations is heavily based on OpenCV, which allows it to beat torchvision, Kornia, PIL, and other similar libraries. But there is still a huge room for improvement in terms of performance, as for some low le
5.
OpenCV 5 Is Here: The Biggest Leap in Years for Computer Vision(opencv.org)
865 points by ternaus 3mo ago|148 comments
6.
by ternaus 4mo ago
It is already there. Submitted 3 weeks ago: https://github.com/pytorch-fdn/ecosystem/issues/67
7.
by ternaus 4mo ago
What is sad is that: - many projects are arrived. - It is unclear who is responsible for the updates. I work on one of the projects in the list, need to update a link to the project, as old one is not actual anymore. And unclear how to do i
8.
by ternaus 5mo ago
Reminds me approach that you get in nearly every book on "How to meet girls". Systematic, efficient. Played this game myself. And I did it when moved to the US with a limited English and lack of understanding of the local culture
9.
by ternaus 7mo ago
I wrote a long practical guide on image augmentation based on ~10 years of training computer vision models and ~7 years maintaining Albumentations. Despite augmentation being used everywhere, most discussions are still very surface-level (“
10.
Ask HN: Where does modern geometry survive contact with SGD?
1 points by ternaus 8mo ago|0 comments
11.
Anthropic Performance Team Take-Home for Dummies(ikot.blog)
1 points by ternaus 8mo ago|0 comments
12.
A String Library Beat OpenCV at Image Processing by 4x(ashvardanian.com)
5 points by ternaus 1y ago|0 comments
13.
by ternaus 1y ago
If standard approach does not work => time to learn how to work hard: [1] For every position you look for people at LinkedIn in that company. If you are already connected ask for referral. [2] If not, look for common connections that can
14.
by ternaus 1y ago
Does it work better than inputing the same information to ChatGPT?
15.
by ternaus 1y ago
On a parallel note. I did not find good engine for the python package documentation, built such an engine myself: https://albumentations.ai/
16.
by ternaus 1y ago
> They wanted it. They paid for it. They enjoyed it. The counter example is open-source software. If we talk about popular packages: - people want it - people enjoy it - people do not pay for that But force-feeding with strict licenses l
17.
by ternaus 1y ago
If there is a strong demand for a feature, regardless of the source of the request - good enough reason to add it.
18.
by ternaus 1y ago
Do it in a very similar way in Cursor: Steps: [1] Add pre-commit hook [2] Write design doc as mdc file to .cursor/rules [3] Iterate on design doc till it describes what I want [4] Ask to write the code [5] Where possible - extend the t
19.
by ternaus 1y ago
I am all for intuitive interfaces, but I am also a big proponent of learning hot keys in every program I work with. It would be better for design to be intuitive, but you struggle only the first time, while interfaces overloaded with inform
20.
by ternaus 1y ago
Interesting representation. Not rows in the database as samples and columns as features, but a whole graph. Makes training much more flexible, and fine tuning as well. Now, when a new data in terms of samples or new tables are connected to
21.
by ternaus 1y ago
Really nice. Useful by itself, but I suspect that will get even more adoption in industry in the telemetry as details statistics about users is ultra valuable.
22.
by ternaus 1y ago
"it is desirable for the software to run as a locally installed executable on your device, rather than a tab in a web browser." An OS agnostic apps, meaning web apps is such a killer feature. You can use the same app on: Linux, Ma
23.
by ternaus 1y ago
I guess, I am experienced open-source developer ( https://github.com/albumentations-team/Albumentations ) 15k stars, 5 million monthly downloads ---- It may happen that Cursor in the agentic mode writes code slower than
24.
by ternaus 1y ago
Not directly related but still. A couple weeks ago I: 1. forked repository of the Albumentations library (15k stars, 5 Million monthly downloads, MIT license) and called it AlbumentationsX 2. changed the license of the fork to the Dual (res
25.
Albumentations: Licensing Change and Project Fork(albumentations.ai)
1 points by ternaus 1y ago|1 comments
26.
by ternaus 1y ago
[1] Open source image augmentations library Albumentations has passed 15k stars on GitHub and reached 94M+ total downloads ( https://clickpy.clickhouse.com/dashboard/albumentations ). [2] Albumentations is no longer main
27.
Ask HN: Need your help with Albumentations feedback
1 points by ternaus 2y ago|0 comments
28.
by ternaus 2y ago
The same, read a few books, all recommended more or less the same. => while was working full time and had income I've sold all stock that got from the employee and put into VTSAX. Living from these money for the past few years in th
29.
Show HN: Leaderboard of Top GitHub Repositories Based on Stars(githublb.vercel.app)
10 points by ternaus 2y ago|1 comments
30.
by ternaus 2y ago
If you speak Russian: https://habr.com/

More ›