SkalskiP

@skalskip92 • 47,635 subscribers

Open-source Lead @roboflow. VLMs. GPU poor. Dog person. Coffee addict. Dyslexic. | GH: https://t.co/dEmzMDGXVf | HF: https://t.co/4Lx1Yw3CLF

Shorts

1,473,978 views

652,911 views

1,572,248 views

708,147 views

810,780 views

72,190 views

79,277 views

54,552 views

592,161 views

504,759 views

121,367 views

445,915 views

197,880 views

336,790 views

241,301 views

112,872 views

72,304 views

165,745 views

49,734 views

Videos

sweetdream.ai

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Private Show

Join now for exclusive access

Free preview available • Premium content

886,012 views • 1 month ago

2,050,207 views • 2 years ago

295,331 views • 8 months ago

329,670 views • 11 months ago

21,213 views • 19 days ago

293,191 views • 1 year ago

427,286 views • 1 year ago

303,581 views • 1 year ago

135,852 views • 7 months ago

72,956 views • 4 months ago

399,903 views • 2 years ago

100,536 views • 6 months ago

266,005 views • 1 year ago

93,025 views • 8 months ago

159,608 views • 1 year ago

This media may contain sensitive content.

96,057 views • 1 year ago

152,451 views • 2 years ago

107,465 views • 1 year ago

188,590 views • 2 years ago

124,358 views • 1 year ago

Live Cam

SkalskiP

Shorts

there's no catch; SAM3 is open source and really good one of the things it does really well is object tracking, even in crazy complex scenes like basketball probably my favorite computer vision model ever

looks like we are working on the same thing haha; more soon

I can finally map NBA player's position from the camera perspective onto the court map it's still a bit shaky... I'll smooth it out later it's time to detect shooting motions and mark the shot location! some of the code has already been migrated to:

SAM3 video tracking is so good yesterday: collect data, train custom object detector, use tracker to estimate object motion - days today: track anything with text prompt - seconds

I used to hate working on projects like this... crazy how the word has changed over the past year; labeling is dead!

SAM 3D Body is a CVPR 2026 award candidate paper from AI at Meta model recovers a full 3D human body mesh from a single RGB image you can run it automatically, or guide the reconstruction with masks and 2D keypoints thx to Niels Rogge for awesome demo idea

RF-DETR is now available in Hugging Face transformers state of the art in both detection and segmentation, outperforming YOLO architectures - checkpoints: - demo: - docs:

INSID3 segments objects across domains using ONLY ONE annotated example it works entirely without a segmentation decoder, task-specific fine-tuning, or external mask generators like SAM CVPR 2026 paper with enormous practical potential

this might be the coolest-looking football AI visualization I ever created

supervision - a computer vision library I created - just crossed 15,000 stars on GitHub! BBBRRRRRR! link:

pip install trackers trackers v2.1.0 is out; this release adds support for ByteTrack, a fast tracking by detection algorithm focused on stable identities under occlusion. link:

I'm starting to get more and more serious with YOLO-World; trying to solve real-life problems. I wanted to see if YOLO-World could recognize that the holes had been filled out. It was pretty tricky, but I learned a little about prompting. ↓ read more

"can you do the same thing for football?" - I got this question 100x today yep, I actually did this last year - player detection and tracking - team clustering - perspective transformation end-to-end YT tutorial:

I'm taking my football/soccer project to the next level today, I worked on detecting players, referees, and the ball and mapping their positions from video frames to positions on the field. ↓ read more

almost fully functional version of my football AI project today, I added player tracking using ByteTrack and projection of players onto the map code coming soon:

supervision-0.26.0 is out we finally released support for ViTPose and ViTPose++ pose estimation models from Hugging Face transformers link:

upcoming basketball AI tutorial will be packed with strong visuals; can't wait to ship it

supervision-0.24.0 is out! you can finally count per-class line crossings. many of you have been asking for this, now we have it! it took me barely 30 minutes to make this demo using supervision! link:

it’s alive. finally trained my first SO-ARM101 policy. basic autonomous pick and place works. more complex stuff soon.

Videos

Watch Anya Live

supervision just hit 40,000 GitHub stars! it now powers over 6.5k open-source computer vision projects, including all my demos like basketball AI link:

football AI code is finally open-source - player detection and tracking - team clustering - camera calibration I still need to work on README; don't judge me on that code:

manual data labeling is over

supervision, the open-source library I created 2 years ago, is crossing 30,000 stars on GitHub! thank you to everyone who helped me build this project! it took us 4,000+ commits, 1,000+ PRs and 100+ contributors to do it. link:

few weeks ago I gave a talk about computer vision and sports 45 minutes, but you can probably watch it at 2x speed over dinner link:

I'm experimenting with using SAM2 for automatic tracking of NBA players; it's mind-blowing how well this model performs out of the box. we might even add a SAM2-based tracker to the trackers package at some point. trackers:

supervision, the open-source library I created a year ago, has crossed 20,000 stars on GitHub this weekend! thank you to everyone who helped me build this project! it took us 3,500+ commits, 850+ PRs and 80+ contributors to do it. repository:

true AGI check your split the G score haha with Roboflow

data labeling is dead. long live distillation. from data to object detection endpoint in 90 seconds. link:

been a minute since I worked with soccer, but trackers-2.3.0 dropped and I had to try OC-SORT it's built for scenes where objects constantly occlude each other, exactly what happens on a soccer pitch. ran it on SoccerNet. looks great. link:

here is the final version of my vehicle speed estimation demo read the thread below to learn how I built it. I will cover: - detection - tracking - perspective transformation - speed calculation - some bonus ideas ↓

check out my mini terminator spent some time over the last few days teaching my reachy mini to follow a face; there is still room for improvement, especially in the tracking algo

FLUX inpainting I know I'm late to the FLUX party, but I hope it's still okey if I drop something cool.

- if SAM3 is so good, do we still need other models? - yes. SAM3 is too large for low compute edge devices and too slow for real-time use. we still need light and fast models for edge workloads. distill SAM3 into smaller model and deploy it in seconds with Roboflow rapid

this guy took my Football AI project and used it to build a 3D representation of the game; CRAAAAAZY!

Sensitive content

CVPR 2025 papers pt. 1 - Gaze-LLE Gaze-LLE simplifies gaze target estimation by building on top of a frozen DINOv2 visual foundation model; SOTA performance; open source code and model more papers: ↓ more

lots of you asked if soccer AI can detect the ball; I just added this functionality code:

this is how you annotate data in 2025 (I've been told this egg demo may trigger some US people)

Automated NBA match commentary using OpenAI vision and TTS (with code!) Everyone is bragging about projects that generate automatic video commentary, but no one is showing the code. I did it while waiting for the plane. code:

SAMURAI: Adapting Segment Anything Model for Zero-Shot Visual Tracking with Motion-Aware check out this SAM2 vs SAMURAI comparison! - paper: - code: - license: Apache-2.0