Ruku

388 posts

Ruku banner
Ruku

Ruku

@lohba_

Building with agents | AI, crypto, financial markets | degen at heart

New York, USA Katılım Aralık 2013
703 Takip Edilen291 Takipçiler
Sabitlenmiş Tweet
Ruku
Ruku@lohba_·
Swapped my ball tracker from TrackNet (3-frame temporal CNN) to RF-DETR (transformer, single-frame). Trained on ~4,700 labeled pickleball frames from public data. Detection rate jumped from 47% to 86%.
English
9
15
249
46.8K
Ruku
Ruku@lohba_·
Before vs after fine-tuning RF-DETR on just 7 labeled frames
Ruku tweet mediaRuku tweet media
English
0
0
1
68
Ruku
Ruku@lohba_·
Fine-tuned RF-DETR on just 7 labeled frames to detect boats in a packed marina from drone footage. 7 felt like way too few to be a real test but it worked surprisingly well.
English
3
0
20
944
Ruku
Ruku@lohba_·
@RyanDiew awesome, thanks for sharing!
English
0
0
1
79
Ryan Diew
Ryan Diew@RyanDiew·
@lohba_ grounding dino auto-labels the frames, then i distill it into a custom yolov11 that runs fast on-device. hard-example mining catches what it misses.
English
1
0
2
199
SkalskiP
SkalskiP@skalskip92·
looks like we are working on the same thing haha; more soon
Google DeepMind@GoogleDeepMind

We’re teaming up @Palmeiras, the first football club to meaningfully build upon TacticAI: our AI system that can help simulate field scenarios and predict open play dynamics up to 8 seconds in advance. ⚽

English
56
179
3.6K
649.1K
Ruku
Ruku@lohba_·
infinite velodrome
Português
0
0
2
389
Ruku
Ruku@lohba_·
8 seconds is enough time for a Polymarket degen to get a real edge in live sports betting
Google DeepMind@GoogleDeepMind

We’re teaming up @Palmeiras, the first football club to meaningfully build upon TacticAI: our AI system that can help simulate field scenarios and predict open play dynamics up to 8 seconds in advance. ⚽

English
0
0
0
199
Ruku
Ruku@lohba_·
Shots from one of my rec. pickleball games visualized in 3D: - Contact timing from audio - Player positions from vision - Ball flight reconstructed between contacts - All from a single iphone video
English
2
1
64
3.7K
Ruku
Ruku@lohba_·
3D body overlay on players with off-the-shelf HMR2.0
English
0
2
26
1.1K
Ruku
Ruku@lohba_·
Adding ball-speed metrics. The ball is tracked, its flight fit to a ballistic arc, and each shot's pace clocked at the bounce. Shown as a blue-to-red meter for relative pace rather than a hard mph.
English
1
1
42
4.8K
Ruku
Ruku@lohba_·
I ran 2D pose estimation on a pickleball rally, with skeletons on all four players through the entire point: - Keypoint R-CNN (torchvision) detects the 17 body joints for each player - PyTorch runs the model inference - OpenCV connects the joints and draws the skeleton overlay onto the frames
English
1
13
91
3.9K
Ruku
Ruku@lohba_·
For each frame, I detect every player and map their feet from pixel space onto real court coordinates using the court homography, then tag each position with the correct player identity. You can accumulate all of a given player's positions into a two-dimensional grid laid over the court and color each cell by how often that player stood there, so that denser areas represent more time spent and the result is a per-player occupancy heatmap.
Ruku tweet media
English
1
5
48
4.9K
Ruku
Ruku@lohba_·
@kirbytheodor Neat, will be following your journey 👍🏻
English
0
0
2
831
Theo
Theo@kirbytheodor·
@lohba_ a home server, mainly for local inference
English
1
0
7
6.1K
Theo
Theo@kirbytheodor·
best investment of my life
Theo tweet mediaTheo tweet mediaTheo tweet media
English
52
15
1.2K
151.4K
uɐɥdǝʇS
uɐɥdǝʇS@StephanSturges·
Every @balajis rant starts like « In order to understand how society is dissolving into a post-marxists neo-authoritarian oligarchy of Bitcoinmaxxing permanently-online uber-chuds fighting the rulemaking post-coldwar overclass you have to consider eggs. Do you know about eggs?»
uɐɥdǝʇS tweet media
English
1
0
4
301
Ruku
Ruku@lohba_·
YOLOv8x detection + ByteTrack tracking (Roboflow Inference & Supervision)
English
0
3
38
3K