Holy f*** I had a single Roboflow tab open somewhere in the background, just showing the project overview, and it ended up consuming FIVE GIGABYTES OF RAM after a day or two. ๐คฏ
What the hell is going on with web development. ๐
Holy f*** I had a single Roboflow tab open somewhere in the background, just showing the project overview, and it ended up consuming FIVE GIGABYTES OF RAM after a day or two. ๐คฏ
What the hell is going on with web development. ๐
Ilir Aliu (@IlirAliu_)
๋ฐ๋๋ฆฌํํธ ์ค ํ๋ฆฌ ๊ตด๊ณก์ ์ธก์ ํ๋ ์ปดํจํฐ ๋น์ ์ฑ ์ฌ๋ก๋ฅผ ์๊ฐํฉ๋๋ค. Roboflow์ RF-DETR๋ก ์ฌ๋์ ์ธ๊ทธ๋ฉํ ์ด์ ํ๊ณ , Ultralytics YOLO11n์ผ๋ก ๋ฐ์ด๋ฉ ๋ฐ์ค๋ฅผ ์์ธกํ๋ ๋ฑ ์ด๋ ์์ธ ๋ถ์์ AI ๋น์ ๊ธฐ์ ์ ํ์ฉํ ์์ฉ์ ๋๋ค.

For all my lifters: computer vision app to measure back curvature during deadlift! main technical highlights: โ RF-DETR (Roboflow) to segment the person (great performance out-the-box with no additional training!) โ YOLO11n (Ultralytics) for bounding box prediction around the

AA (@measure_plan)
Gemma 4์ ๋น์ ๊ธฐ๋ฅ์ ํ์ฉํด Roboflow RF-DETR๋ก ๊ฐ์ฒด๋ฅผ ๋จผ์ ํ์งํ๊ณ , Gemma๊ฐ ์ฅ๋ฉด์ ํ ๋ฌธ์ฅ์ผ๋ก ์์ฝํ๋ ์ฑ์ ๋ง๋ค์ด ์คํํ์ต๋๋ค. ๊ฐ์ฒด ์ธ์๊ณผ ๋ฉํฐ๋ชจ๋ฌ ์์ฝ์ ๊ฒฐํฉํ ์ค์ฉ์ AI ์์ฉ ์ฌ๋ก์ ๋๋ค.

i spent the afternoon experimenting with Gemma 4's vision capabilities made an app that uses roboflow RF-DETR for a first pass of object detections and Gemma to summarize the scene in one sentence for fun i asked Gemma to "describe what you see as if you were a medieval bard"
AA (@measure_plan)
์๋ฐ์คํฌ๋ฆฝํธ, ์บ๋ฒ์ค, Roboflow RF-DETR, MediaPipe ํฌ์ฆ ํธ๋ํน์ ํ์ฉํด ํธ์์ ๊ณผ ํ๋ญํฌ๋ก ์กฐ์ํ๋ ๋ํค์ฝฉ ์คํ์ผ ํผํธ๋์ค ๊ฒ์์ ๋ง๋ค์๋ค. AI ๋น์ ๊ณผ ์ฒด์ฑ ์ธ์ ๊ธฐ์ ์ ๊ฒฐํฉํ ํฅ๋ฏธ๋ก์ด ์ธํฐ๋ํฐ๋ธ ํ๋ก์ ํธ๋ค.

i made donkey kong but you play with pushups and planks and it's quite inconvenient made with javascript, canvas, roboflow RF-DETR (for banana detection), and mediapipe body pose tracking lmk if you want to play a few fitness games like this and i'll try to get these online
AA (@measure_plan)
์ฌ์ง ๊ฒ์ยท๋ทฐ์ ์คํ: threejs ์ธํฐํ์ด์ค ์์ smolVLM + Roboflow RF-DETR ๊ธฐ๋ฐ ์ปดํจํฐ ๋น์ ๊ณผ ChromaDB ์๋งจํฑ ๊ฒ์์ ๊ฒฐํฉํด, ๋์ยท์์ยท๊ฐ์ฒด ๊ฐ์ ์์ฐ์ด๋ก ์ ์ฒด ์ฌ์ง ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ๊ฒ์ํ ์ ์๊ฒ ๊ตฌ์ฑํ ์ฌ๋ก๋ฅผ ๊ณต์ ํจ.
AA (@measure_plan)
๋ก์ปฌ์ ๋ฌด๋ฃ AI ๋ชจ๋ธ๋ค์ ํ์ฉํด ์ปดํจํฐ์ ์๋ ์ค๋๋ ์ฌํ ์ฌ์ง์ ์์ฐ์ด๋ก ๊ฒ์ํ๋ ์ฑ์ ์ ์ํ ์ฌ๋ก: smolVLM์ผ๋ก ์ฅ๋ฉดยท์์ ์ค๋ช , Roboflow RF-DETR๋ก ๊ฐ์ฒด ๊ฒ์ถ, ChromaDB์ ๋ฉํ๋ฐ์ดํฐ ์ ์ฅยท์๋งจํฑ ๊ฒ์์ ์ํํ๊ณ Python๊ณผ Streamlit์ผ๋ก UI๋ฅผ ๊ตฌํํ์ต๋๋ค.

i built an app to search old travel photos on my computer with natural language queries using free local AI models: - smolVLM to describe the scene and colours - roboflow RF-DETR to detect objects - chromaDB to store metadata labels and run semantic search - python + streamlit
AA (@measure_plan)
'fruit xylophone'๋ผ๋ ์ปดํจํฐ ๋น์ ๊ธฐ๋ฐ ์์ ์คํ์ ์๊ฐํ๋ ํธ์์ผ๋ก, mediapipe ์ ์ถ์ ๊ณผ Roboflow์ RF-DETR ๊ฐ์ฒด ๊ฒ์ถ์ ๊ฒฐํฉํด ๋ธ๋ผ์ฐ์ ์์ ์ค์๊ฐ์ผ๋ก ๋์ํ๋ฉฐ transformers.js๋ฅผ ์ฌ์ฉํด ๊ตฌํํ๋ค๊ณ ์ค๋ช ํฉ๋๋ค. ์ค์๊ฐ ๋ธ๋ผ์ฐ์ ML๊ณผ ํฌ๋ฆฌ์์ดํฐ๋ธ ์ธํฐ๋์ ์ฌ๋ก์ ๋๋ค.
AA (@measure_plan)
smolVLM, transformers.js, RF-DETR ๋ชจ๋ธ ๋ฑ ๋ค์ํ ์คํ์์ค ๋น์ ๊ด๋ จ ํ๋ก์ ํธ๊ฐ ์๊ฐ๋์์ต๋๋ค. Hugging Face์ Roboflow ํ์ด ์ฃผ๋ํ ์ด ๋๊ตฌ๋ค์ ์ปดํจํฐ ๋น์ ํ์ต์๋ค์๊ฒ ์ ์ฉํ ์ค์ต ์์ ์ ๋ฐ๋ชจ๋ฅผ ์ ๊ณตํฉ๋๋ค.

thank you to @xenovacom, @mervenoyann, @ngxson and the hugging face team for smolVLM, transformers js, and all the blog posts + demos thank you to @skalskip92 and the roboflow team for the RF-DETR model if you want to learn computer vision, follow these folks and their awesome