๐Ÿ“ข ๊ฒ€์ƒ‰ ๊ธฐ๋Šฅ ์ถ”๊ฐ€ ์˜ˆ์ •

์ฃผ๋ฏผ๊ฑด

๊ธฐ๋ก = ์„ฑ์žฅ

zoomg

23.06.14 (Wed)

Open LLM Leaderboard - a Hugging Face Space by HuggingFaceH4
Discover amazing ML apps made by the community
I-JEPA: The first AI model based on Yann LeCunโ€™s vision for more human-like AI
I-JEPA learns by creating an internal model of the outside world, which compares abstract representations of images (rather than comparing the pixels themselves).
without having the model fill in pixel-level details, which tend to result in learning less semantically meaningful representations.
Segment Anything: TorchServe๋กœ ๋ฐฐํฌํ•˜๊ธฐ - code ํฌํ•จ
Segment Anything model์„ TorchServe๋ฅผ ํ†ตํ•ด ๋ฐฐํฌํ•˜๋Š” ๊ณผ์ •์„ Code๋ฅผ ํ†ตํ•ด์„œ ์ƒ์„ธํ•˜๊ฒŒ ์•Œ๋ ค๋“œ๋ฆฝ๋‹ˆ๋‹ค. python code์™€ docker๋ฅผ ์‚ฌ์šฉํ•ด ์ถ”๋ก  ํ™˜๊ฒฝ์„ ๋งŒ๋“œ๋Š” ๊ณผ์ •์ด ํฌํ•จ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค.
๋…ผ๋ฌธ๋ฆฌ๋ทฐ Segment Anything | SAM
Segment Anything Model์€ SAM์€ ๊ฐ์ฒด ์‹๋ณ„ ์„ฑ๋Šฅ, ํด๋ž˜์Šค(=๊ฐ์ฒด ์นดํ…Œ๊ณ ๋ฆฌ) ์ž์œ ๋„, ์ž‘๊ฑฐ๋‚˜ ๋ณต์žกํ•œ ๊ฐ์ฒด์— ๋Œ€ํ•œ ์„ธ๋ถ„ํ™”๋œ ๋ถ„ํ• , ์‹ค์‹œ๊ฐ„ ์ฒ˜๋ฆฌ, ์‰ฌ์šด ์„œ๋น„์Šค ํ†ตํ•ฉ ๋ฐ ์‚ฌ์šฉ ๋“ฑ ๋‹ค์–‘ํ•œ ์ธก๋ฉด์—์„œ ์ด์ ์„ ๊ฐ–๋Š” ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค.
Comparing current and prior imaging studies in radiology AI
Todayโ€™s AI radiology solutions donโ€™t effectively account for past imaging findings despite their routine use by radiologists. BioViL-T jointly trains text and image encoders that can ground radiology text across longitudinal images:

23.06.08 (Thu)

๋”ฅ๋ธŒ๋ ˆ์ธ, ์ƒˆ๋กœ์šด AI ์ •๋ ฌ ์•Œ๊ณ ๋ฆฌ์ฆ˜ ๋ฐœ๊ฒฌ...10์—ฌ๋…„ ๋งŒ์˜ ์ง„์ „
๋”ฅ๋งˆ์ธ๋“œ๊ฐ€ ์ƒˆ๋กœ์šด AI โ€˜์•ŒํŒŒ๋ฐ๋ธŒ(AlphaDev)โ€™๋ฅผ ์‚ฌ์šฉํ•ด ๋” ๋น ๋ฅธ ์ •๋ ฌ ์•Œ๊ณ ๋ฆฌ์ฆ˜(sorting algorithm)์„ ๋ฐœ๊ฒฌํ–ˆ๋‹ค. AI๊ฐ€ ์ „์— ์—†๋˜ ์ƒˆ๋กœ์šด ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ๋ฐœ๊ฒฌํ•ด ์ธ๊ฐ„์˜ ๊ธฐ์ˆ  ์ˆ˜์ค€์„ ๋›ฐ์–ด๋„˜์„ ์ˆ˜ ์žˆ๋‹ค๋Š” ๊ฒƒ์„ ๋ณด์—ฌ์ค€ ์ ์—์„œ ์ฃผ๋ชฉ๋œ๋‹ค.๋”ฅ๋งˆ์ธ๋“œ ์—ฐ๊ตฌ์ง„์€ 7์ผ(ํ˜„์ง€์‹œ๊ฐ„) ๋„ค์ด์ฒ˜์— ๊ฐ•ํ™”ํ•™์Šต ๊ธฐ๋ฐ˜์˜ ์ƒˆ๋กœ์šด AI โ€˜์•ŒํŒŒ๋ฐ๋ธŒโ€™๊ฐ€ ๊ธฐ์กด๋ณด๋‹ค ๋น ๋ฅด๊ณ  ํšจ์œจ์ ์ธ ์ƒˆ ์ •๋ ฌ ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ๋ฐœ๊ฒฌํ–ˆ๋‹ค๊ณ  ๋ฐœํ‘œํ–ˆ๋‹ค.์—ฐ๊ตฌ์ง„์€ ์•ŒํŒŒ๋ฐ๋ธŒ๊ฐ€ ๋ฐœ๊ฒฌํ•œ ์ƒˆ ์ •๋ ฌ ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ํ”„๋กœ๊ทธ๋ž˜๋ฐ ์–ธ์–ด์ธ C++์˜ ํ‘œ์ค€ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์— ๊ธฐ๋ณธ ์•Œ๊ณ ๋ฆฌ์ฆ˜์œผ๋กœ ์ถ”๊ฐ€ํ–ˆ๋‹ค. ์ง€๋‚œ 10์—ฌ๋…„ ๋™์•ˆ ์ƒˆ

23.06.06 (Tue)

์ƒˆ๋กœ์šด ์ปดํ“จํŒ… ํ™˜๊ฒฝ์„ ์ œ์•ˆ, ๋‹ค๋ฅธ ์—…์ฒด๋“ค๊ณผ ๋‹ฌ๋ฆฌ ๋ช…ํ™•ํ•œ ๋ชฉํ‘œ์— ์˜ํ•ด ์„ค๊ณ„๋œ ๋А๋‚Œ์ด ๊ฐ•ํ•จ, ์ œํ’ˆ์„ ์†Œ๊ฐœํ•˜๋Š” ์˜์ƒ์—์„œ๋ถ€ํ„ฐ ์–ด๋–ค ๋ชฉํ‘œ๋ฅผ ๊ฐ€์ง€๊ณ  ์ œํ’ˆ์„ ๊ฐœ๋ฐœํ–ˆ๋Š”์ง€ ์•Œ ์ˆ˜ ์žˆ์Œ
OpenAIโ€™s plans according to Sam Altman
Last week I had the privilege to sit down with Sam Altman and 20 other developers to discuss OpenAIโ€™s APIs and their product plans. Sam was remarkably open. The discussion touched on practical developer issues as well as bigger-picture questions related to OpenAIโ€™s mission and the societal impact ofโ€ฆ

Great! Youโ€™ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to zoomg.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.