RE: https://mastodon.social/@xlth/116144192667591833

Not the ideal conditions for geospatial applications of VLMs ๐Ÿ˜…

#GIScience #VLM #spatiotemporal #MobilityDataScience #SpatialDataScience

็”ฐไธญ็พฉๅผ˜ | taziku CEO / AI ร— Creative (@taziku_co)

๋กœ๋ด‡๊ณผ VLM(๋น„์ „-์–ธ์–ด ๋ชจ๋ธ)์„ ์‹ค์ œ ๋™์ž‘ ๋ฃจํ”„(๋ณด๋Š”โ†’ํŒ๋‹จโ†’์›€์ง์ด๋Š”โ†’๋‹ค์‹œ ๋ณด๋Š”) ๊ด€์ ์—์„œ ์ธก์ •ํ•˜๋Š” ํ”Œ๋žซํผ์ธ AgenticLab์„ ์†Œ๊ฐœํ•˜๋Š” ๋‚ด์šฉ์ž…๋‹ˆ๋‹ค. ๊ธฐ์กด VQA ๊ฐ™์€ ์ •์  ๋ฒค์น˜๋งˆํฌ์™€ ๋‹ฌ๋ฆฌ ์‹ค์ œ ๋กœ๋ด‡์—์„œ ์—ฐ์†์ ์ธ ์ธ์ง€-ํ–‰๋™ ๋ฃจํ”„๋ฅผ ํ‰๊ฐ€ํ•˜๋Š” ์ ์„ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค.

https://x.com/taziku_co/status/2029419731258843415

#robotics #vlm #vqa #agenticlab

็”ฐไธญ็พฉๅผ˜ | taziku CEO / AI ร— Creative (@taziku_co) on X

ใƒญใƒœใƒƒใƒˆร—VLMใ‚’็œŸๆญฃ้ขใ‹ใ‚‰ๆธฌใ‚‹ ใ€Œ่ฆ‹ใ‚‹AIใ€ใฏVQAใชใฉใฎใƒ™ใƒณใƒใƒžใƒผใ‚ฏใง่ฉ•ไพกใ•ใ‚Œใฆใ„ใ‚‹ใ€‚ ใงใ‚‚ใƒญใƒœใƒƒใƒˆใงใฏ ่ฆ‹ใฆโ†’ๅˆคๆ–ญโ†’ๅ‹•ใโ†’ใพใŸ่ฆ‹ใ‚‹ ใ“ใฎใƒซใƒผใƒ—ใŒใšใฃใจ็ถšใใ€‚ AgenticLabใฏๅฎŸๆฉŸใƒญใƒœใƒƒใƒˆใงใ“ใฎไธ€้€ฃใฎๅ‹•ไฝœใ‚’่ฉ•ไพกใ™ใ‚‹ใƒ—ใƒฉใƒƒใƒˆใƒ•ใ‚ฉใƒผใƒ ใ€‚ ่ฉณ็ดฐใฏ๐Ÿงต

X (formerly Twitter)

้‡‘ใฎใƒ‹ใƒฏใƒˆใƒช (@gosrum)

๋กœ์ปฌ VLM(๋น„์ฃผ์–ผ ์–ธ์–ด ๋ชจ๋ธ)์€ Mac์ด ์•„๋‹ˆ๋ผ NVIDIA GPU์—์„œ ์ถ”๋ก ํ•˜๋ฉด ์‹ค์‹œ๊ฐ„ ์ฒ˜๋ฆฌ๊ฐ€ ๊ฐ€๋Šฅํ•  ๊ฒƒ ๊ฐ™๋‹ค๋Š” ๊ฐœ์ธ ์‹คํ—˜ ๊ธฐ๋ก์ž…๋‹ˆ๋‹ค. ์ด๋ฅผ ํ™•์ธํ•˜๋ ค๊ณ  Ubuntu(Linux)์—์„œ Claude Code๋ฅผ ์‚ฌ์šฉํ•ด ์•ฑ์„ ์žฌํ˜„ํ•ด๋ดค์ง€๋งŒ Wayland ํ™˜๊ฒฝ์—์„œ ๊ตฌํ˜„์ด ์–ด๋ ค์› ๋‹ค๋Š” ๊ฒฝํ—˜์„ ๊ณต์œ ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.

https://x.com/gosrum/status/2028426240470999164

#vlm #nvidia #claude #ubuntu #wayland

้‡‘ใฎใƒ‹ใƒฏใƒˆใƒช (@gosrum) on X

ใƒญใƒผใ‚ซใƒซVLMใฏMacใ˜ใ‚ƒใชใใฆNVIDIA GPUใงๆŽจ่ซ–ใ™ใ‚Œใฐใƒชใ‚ขใƒซใ‚ฟใ‚คใƒ ใงใ„ใ‘ใ‚‹ใ‚“ใ˜ใ‚ƒใชใ„ใ‹ใจๆ€ใฃใฆใ„ใ‚‹ ใใ‚Œใ‚’่ฉฆใ™ใŸใ‚ใซใ€Linux๏ผˆUbuntu๏ผ‰ใงใ“ใฎใ‚ขใƒ—ใƒชใ‚’ๅ†็พใ™ใ‚‹ในใClaude Codeไฝฟใฃใฆ่ฉฆใ—ใฆใŸใ‘ใฉใ€ใ‚„ใฏใ‚ŠWaylandใ ใจ้›ฃใ—ใใ†

X (formerly Twitter)

merve (@mervenoyann)

๋น„์ „-์–ธ์–ด ๋ชจ๋ธ(VLM) ๊ด€๋ จ ์„œ์ ์— ์ƒˆ๋กœ ๋‘ ์žฅ์ด ์ถ”๊ฐ€๋˜์—ˆ๋‹ค๋Š” ๊ณต์ง€์ž…๋‹ˆ๋‹ค. ๋ฌธ์„œ AI ์žฅ์€ ๊ธฐ์กด ๋ชจ๋ธ๋ถ€ํ„ฐ ์ตœ์‹  VLM ์ ‘๊ทผ๋ฒ•, ๊ฒ€์ƒ‰(retrieval) ๋“ฑ์„ ์ •๋ฆฌํ•˜๊ณ , ๋น„๋””์˜ค ์–ธ์–ด ๋ชจ๋ธ ์žฅ์€ ๋น„๋””์˜ค ์ดํ•ด์™€ ๊ด€๋ จ ๊ธฐ๋ฒ• ๋ฐ ์‹ค๋ฌด ๋…ธํ•˜์šฐ๋ฅผ ๋‹ค๋ฃน๋‹ˆ๋‹ค. ๋ฐœํ‘œ์ž๊ฐ€ ์ง์ ‘ ์ง‘ํ•„ํ•œ ๋ฌธ์„œ AI ์žฅ๋„ ํฌํ•จ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค.

https://x.com/mervenoyann/status/2028404451476705295

#visionlanguagemodels #documentai #videolanguagemodels #vlm

merve (@mervenoyann) on X

two more chapters on vision language models book is out! > document AI chapter (by yours truly) shows old models, new VLM approaches, retrieval and more! > video language models chapter shows video understanding, know-hows, approaches and more! sneak peek below

X (formerly Twitter)

Manchester Monday 5 23rd February 2026.

OO-VLF, Fokker F50, VLM, flying past the Runway Visitor Park shortly after taking off from Runway 23 Right at Manchester Airport, 24th August 2015.

#Manchester #MAN #EGCC #Runway23Right #RunwayVisitorPark #Fokker #F50 #VLM #ManchesterMonday
#AvGeek #aviation #planespotting #photography

Manchester Monday 23rd February 2026.

[โ€ฆ]

https://mancavgeek.co.uk/2026/02/23/manchester-monday-23rd-february-2026/

yzma 1.10 is out with improvements like:
- install info for @officialarduino.bsky.social UNO Q and @raspberrypi.com
- experimental 'VLM' type
- improved yzma cmd so 'go install' works with latest

Go and get it!

https://github.com/hybridgroup/yzma

#golang #llama #llamacpp #ml #llm #vlm #cuda #vulkan

"Running local LLMs and VLMs on the Arduino UNO Q with yzma"

@golang running on the @arduino #unoq can be your new tiny edge inference device!

https://projecthub.arduino.cc/marc-edgeimpulse/running-local-llms-and-vlms-on-the-arduino-uno-q-with-yzma-74e288

#golang #yzma #llama #llamacpp #llm #vlm #arduino #ml

Running local LLMs and VLMs on the Arduino UNO Q with yzma

Discover how to run local LLMs and VLMs directly on the Arduino UNO Q using Ron Evans' yzma project, using llama.cpp with Go making edge AI LLM possible in the Arduino UNO Q.

Arduino Project Hub

Hasan Toor (@hasantoxr)

์•Œ๋ฆฌ๋ฐ”๋ฐ”๊ฐ€ ์ž์‚ฌ ํ”Œ๋ž˜๊ทธ์‹ญ ๋ชจ๋ธ Qwen์˜ ์‹ ๋ฒ„์ „ Qwen3.5๋ฅผ ๊ณต๊ฐœํ–ˆ์Šต๋‹ˆ๋‹ค. Qwen3.5๋Š” ๋„ค์ดํ‹ฐ๋ธŒ ๋ฉ€ํ‹ฐ๋ชจ๋‹ฌ(natively multimodal) ์˜คํ”ˆ-์›จ์ดํŠธ VLM์œผ๋กœ GUI ์ƒํ˜ธ์ž‘์šฉ, ๋น„๋””์˜ค ์ดํ•ด, ์—์ด์ „ํŠธํ˜• ์›Œํฌํ”Œ๋กœ์šฐ์—์„œ ์„ฑ๋Šฅ๊ณผ ํšจ์œจ์„ฑ์„ ์žฌ์ •์˜ํ•˜๋„๋ก ์„ค๊ณ„๋˜์—ˆ์œผ๋ฉฐ ์ฐจ์„ธ๋Œ€ ์• ํ”Œ๋ฆฌ์ผ€์ด์…˜์„ ๊ฒจ๋ƒฅํ•œ ์˜คํ”ˆ ์†Œ์Šค/์˜คํ”ˆ ๊ฐ€์ค‘์น˜ ๋ชจ๋ธ๋กœ ์ฃผ๋ชฉ๋ฉ๋‹ˆ๋‹ค.

https://x.com/hasantoxr/status/2024083448735121748

#alibaba #qwen #vlm #multimodal #openweight

Hasan Toor (@hasantoxr) on X

๐ŸšจBIG NEWS from Alibaba: Qwen's new flagship model Qwen3.5 is here, a natively multimodal, open-weight VLM that redefines efficiency vs. performance. Excels at GUI interaction, video comprehension, and agentic workflows. It's not just smarterโ€ฆ it's built for next-gen

X (formerly Twitter)