Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

LFM2.5-VL-3B recognized Minecraft Steve running locally on an iPhone in Atomic Chat! Liquid AI released a new 3B vision model. We gave it one photo of a small Steve toy and it named the character correctly. At Q4 it takes roughly 2.3 GB. Available now on iPhone and Android!

atomic.chat

14,898 subscribers

10,594 görüntüleme • 1 ay önce •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

DeepSeek V4 Flash + LFM2.5-VL-3B as the vision auxiliary in Hermes Agent is a deadly combo. DeepSeek handles the agentic work. LFM2.5-VL-3B gives it the eyes. I told it to use only its local vision model to browse. It read Hacker News, picked the top story, opened the eclipse map, found an Iceland webcam marker and opened the live feed. Every step seen through a 3B model running locally. No cloud vision API.

DeepSeek V4 Flash + LFM2.5-VL-3B as the vision auxiliary in Hermes Agent is a deadly combo. DeepSeek handles the agentic work. LFM2.5-VL-3B gives it the eyes. I told it to use only its local vision model to browse. It read Hacker News, picked the top story, opened the eclipse map, found an Iceland webcam marker and opened the live feed. Every step seen through a 3B model running locally. No cloud vision API.

Noctus

67,801 görüntüleme • 1 ay önce

Check out our demos using LFM2.5-VL-3B, our latest lightweight, vision-language model that reads screens, documents, and the physical world. First up: LFM2.5-VL-3B running fully on-device in the browser with WebGPU to understand a document page. The model parses the entire layout in one pass and returns regions and labels that the interface renders as an overlay. The demo highlights OCR and layout understanding for visually structured content such as forms, reports, receipts, and other documents. 🧵

Check out our demos using LFM2.5-VL-3B, our latest lightweight, vision-language model that reads screens, documents, and the physical world. First up: LFM2.5-VL-3B running fully on-device in the browser with WebGPU to understand a document page. The model parses the entire layout in one pass and returns regions and labels that the interface renders as an overlay. The demo highlights OCR and layout understanding for visually structured content such as forms, reports, receipts, and other documents. 🧵

Liquid AI

12,613 görüntüleme • 1 ay önce

Advice for AI engineers 💡 A small Visual Language Model fine-tuned on your custom dataset is as accurate as GPT-5... ... and costs 50 times less. For example, LFM2.5-VL-1.6B by Liquid AI running locally at full speed with llama.cpp ↓

Advice for AI engineers 💡 A small Visual Language Model fine-tuned on your custom dataset is as accurate as GPT-5... ... and costs 50 times less. For example, LFM2.5-VL-1.6B by Liquid AI running locally at full speed with llama.cpp ↓

Pau Labarta Bajo

80,352 görüntüleme • 8 ay önce

WeiboAI dropped VibeThinker-3B, so I had to try it locally. this is a 3B model, not a giant frontier system. in the video I load it on my DGX Spark, give it 3 small algorithm problems, stream the actual model output live, then run the generated python through pytest. no benchmark screenshot no canned answer just a tiny local reasoner writing code and real tests deciding if it worked it went 3/3.

WeiboAI dropped VibeThinker-3B, so I had to try it locally. this is a 3B model, not a giant frontier system. in the video I load it on my DGX Spark, give it 3 small algorithm problems, stream the actual model output live, then run the generated python through pytest. no benchmark screenshot no canned answer just a tiny local reasoner writing code and real tests deciding if it worked it went 3/3.

Joey

16,865 görüntüleme • 3 ay önce

DeepSeek V4.1 Flash made DOOM in Atomic Chat 🐳 DeepSeek dropped a new architecture for the flash model so we tested it by making Doom in an html file and it built a real raycasting shooter with gun mechanics and enemy waves Run AI models locally ->

DeepSeek V4.1 Flash made DOOM in Atomic Chat 🐳 DeepSeek dropped a new architecture for the flash model so we tested it by making Doom in an html file and it built a real raycasting shooter with gun mechanics and enemy waves Run AI models locally ->

atomic.chat

29,435 görüntüleme • 1 ay önce

while we don’t know for sure if the iPhone mirroring feature will come to the Apple Vision Pro, I sure hope it does. for now, it's available on Mac—i tested it and it's amazing! here’s how you can experience it on your Mac today: 1. make sure both your iPhone and Mac are running the latest beta versions of iOS 18 and macOS Sequoia. 2. position your iPhone near your Mac. 3. an icon titled "iPhone Mirroring" will appear on your Mac's Dock. 4. select the "Continue" button on your Mac. 5. unlock your iPhone when prompted. 6. on your Mac, click the "Get Started" button. 7. with your iPhone locked, open the "iPhone Mirroring" app on your Mac. this feature is a game-changer, and i can’t wait for the day it hopefully comes to the Vision Pro!

while we don’t know for sure if the iPhone mirroring feature will come to the Apple Vision Pro, I sure hope it does. for now, it's available on Mac—i tested it and it's amazing! here’s how you can experience it on your Mac today: 1. make sure both your iPhone and Mac are running the latest beta versions of iOS 18 and macOS Sequoia. 2. position your iPhone near your Mac. 3. an icon titled "iPhone Mirroring" will appear on your Mac's Dock. 4. select the "Continue" button on your Mac. 5. unlock your iPhone when prompted. 6. on your Mac, click the "Get Started" button. 7. with your iPhone locked, open the "iPhone Mirroring" app on your Mac. this feature is a game-changer, and i can’t wait for the day it hopefully comes to the Vision Pro!

Justin Ryan

70,266 görüntüleme • 2 yıl önce

NEW: Mistral AI releases Mistral 3, a family of multimodal models, including three start-of-the-art dense models (3B, 8B, and 14B) and Mistral Large 3 (675B, 41B active). All Apache 2.0! 🤗 Surprisingly, the 3B is small enough to run 100% locally in your browser on WebGPU! 🤯

NEW: Mistral AI releases Mistral 3, a family of multimodal models, including three start-of-the-art dense models (3B, 8B, and 14B) and Mistral Large 3 (675B, 41B active). All Apache 2.0! 🤗 Surprisingly, the 3B is small enough to run 100% locally in your browser on WebGPU! 🤯

Xenova

225,317 görüntüleme • 10 ay önce

love jev, but upset it - isn't open source? - it lack vision capability? We fixed all of that, introducing A fully open source library which takes any HF model and Jev-ify it, with an API endpoint Now on github, and live in production at Featherless AI

love jev, but upset it - isn't open source? - it lack vision capability? We fixed all of that, introducing A fully open source library which takes any HF model and Jev-ify it, with an API endpoint Now on github, and live in production at Featherless AI

Eugene Cheah - AI builder @ 🇸🇬|🇺🇸

106,163 görüntüleme • 22 gün önce

"Tell me about the sales actuals." It opened the folder, read the spreadsheet, and broke down Q1 revenue by region. A 2.6B model. 2.3 GB on disk. Running in Osaurus, day one.

"Tell me about the sales actuals." It opened the folder, read the spreadsheet, and broke down Q1 revenue by region. A 2.6B model. 2.3 GB on disk. Running in Osaurus, day one.

Osaurus

34,708 görüntüleme • 2 ay önce

Liquid AI's head of post-training explained how they built a small model that runs on-device under 1 GB in 20 minutes - better than $2500 small-model bootcamps. pick LFM2.5 base -> on-policy preference alignment -> agentic reinforcement learning -> curriculum training -> iterative model merging -> ship a 1B model that reliably calls tools on your phone. That loop is why frontier small models now beat 70B models on the tasks that actually matter. LFM2.5 + on-policy DPO + agentic RL + curriculum training + iterative merging - that's the stack. Watch and save it, then run a 1B agent on your phone tonight.

Liquid AI's head of post-training explained how they built a small model that runs on-device under 1 GB in 20 minutes - better than $2500 small-model bootcamps. pick LFM2.5 base -> on-policy preference alignment -> agentic reinforcement learning -> curriculum training -> iterative model merging -> ship a 1B model that reliably calls tools on your phone. That loop is why frontier small models now beat 70B models on the tasks that actually matter. LFM2.5 + on-policy DPO + agentic RL + curriculum training + iterative merging - that's the stack. Watch and save it, then run a 1B agent on your phone tonight.

h100envy

87,975 görüntüleme • 2 ay önce

First demo: The Minecraft Test. Kaigen Engine works great with AI. The difference with Three.js is that when you're done it runs on Windows, Mac, iPhone, Android *and* the Web. (VR soon!) oh and it runs at 330+ FPS on a 6 year old laptop gpu. link to demo + code below🧵

First demo: The Minecraft Test. Kaigen Engine works great with AI. The difference with Three.js is that when you're done it runs on Windows, Mac, iPhone, Android and the Web. (VR soon!) oh and it runs at 330+ FPS on a 6 year old laptop gpu. link to demo + code below🧵

Gabriel Dechichi

37,860 görüntüleme • 16 gün önce

New iPhone 16 Pro—New Network. And no, your hands don’t have to be this gorgeous to hold it. 😉 We’re thrilled to offer the brand new iPhone 16 Pro. Get it on us, no-trade in needed and the freedom to upgrade every year. Available now at #BoostMobile

Boost Mobile

19,597 görüntüleme • 2 yıl önce

Mark Zuckerberg says Apple haven't invented anything great in a while: "Steve Jobs invented the iPhone and now they're just sitting on it 20 years later... so how are they making more money as a company? They do it by squeezing people"

Mark Zuckerberg says Apple haven't invented anything great in a while: "Steve Jobs invented the iPhone and now they're just sitting on it 20 years later... so how are they making more money as a company? They do it by squeezing people"

Tsarathustra

13,470,892 görüntüleme • 1 yıl önce

Entering the #GAMDOM50KCHALLENGE I put everything Steve Will Do It and CODE “steve” on GAMDOM might like all in one piece with a banger song. (We Ricky and Morty-fied Steve and Adukes 😉, and the list follow.) I’m under code steve ID: Ningygmit Steve #1 Gamdom #2

Entering the #GAMDOM50KCHALLENGE I put everything Steve Will Do It and CODE “steve” on GAMDOM might like all in one piece with a banger song. (We Ricky and Morty-fied Steve and Adukes 😉, and the list follow.) I’m under code steve ID: Ningygmit Steve #1 Gamdom #2

Ningy

163,636 görüntüleme • 23 gün önce

NEW: LiquidAI just released LFM2.5-350M, a tiny model that brings agentic AI and tool-calling capabilities to resource-constrained environments. 🤯 It can even run locally in your browser via WebGPU, serving as a powerful companion while you browse the web. Try the demo! 👇

NEW: LiquidAI just released LFM2.5-350M, a tiny model that brings agentic AI and tool-calling capabilities to resource-constrained environments. 🤯 It can even run locally in your browser via WebGPU, serving as a powerful companion while you browse the web. Try the demo! 👇

Xenova

26,267 görüntüleme • 6 ay önce

Most heavy AI users pay $400-500/month in subscriptions. $200 ChatGPT, $200 Claude Code, $20 Cursor, $5,000 a year. He put a small box on his desk, two commands, free model running locally. Same chat, same answers, nothing leaves the room. Paid itself off in 4 months, now costs $3 in electricity. Most people look at this and see a workstation. one group looks at it and stops paying.

Most heavy AI users pay $400-500/month in subscriptions. $200 ChatGPT, $200 Claude Code, $20 Cursor, $5,000 a year. He put a small box on his desk, two commands, free model running locally. Same chat, same answers, nothing leaves the room. Paid itself off in 4 months, now costs $3 in electricity. Most people look at this and see a workstation. one group looks at it and stops paying.

ToraX

24,497 görüntüleme • 3 ay önce

Mark Zuckerberg of $META says Apple, $AAPL, haven't invented anything great in a while: "Steve Jobs invented the iPhone and now they're just sitting on it 20 years later... so how are they making more money as a company? They do it by squeezing people," per tsarnick:

Mark Zuckerberg of $META says Apple, $AAPL, haven't invented anything great in a while: "Steve Jobs invented the iPhone and now they're just sitting on it 20 years later... so how are they making more money as a company? They do it by squeezing people," per tsarnick:

unusual_whales

545,851 görüntüleme • 1 yıl önce

“Between me and the buyer, who is at fault? The buyer is now requesting for another phone.”🤯💔 -Man selling an iPhone ask after a customer’s new iPhone got brøken when it fell on the floor during unboxing.

“Between me and the buyer, who is at fault? The buyer is now requesting for another phone.”🤯💔 -Man selling an iPhone ask after a customer’s new iPhone got brøken when it fell on the floor during unboxing.

CHUKS 🍥

191,614 görüntüleme • 5 ay önce

🚨 In October 2018, Apple released one of its most controversial and criticized iPhones: the iPhone XR. Many dismissed it at launch, but history told a different story. The iPhone XR became one of Apple’s best-selling iPhones As of September 2020, The iPhone XR had sold 77+ million units.

🚨 In October 2018, Apple released one of its most controversial and criticized iPhones: the iPhone XR. Many dismissed it at launch, but history told a different story. The iPhone XR became one of Apple’s best-selling iPhones As of September 2020, The iPhone XR had sold 77+ million units.

Jason C.

82,079 görüntüleme • 2 ay önce