正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Gemma 4 just dropped. I had it captioning video in real-time within an hour. Running locally on a MacBook. No cloud. No API. Real-time scene understanding. Oh and SAM3 is segmenting every object in the same frame. Same laptop.

Maziyar PANAHI

17,371 subscribers

196,846 次观看 • 4 个月前 •via X (Twitter)

教育科学技术

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

🚨 BREAKING: Apple just acquired Meta's SAM team! On-device segmentation now ships natively in macOS 26.x This is what it looks like when you run SAM3 on MLX locally. Every car. Every fire. Real-time. M2 laptop. No cloud.

🚨 BREAKING: Apple just acquired Meta's SAM team! On-device segmentation now ships natively in macOS 26.x This is what it looks like when you run SAM3 on MLX locally. Every car. Every fire. Real-time. M2 laptop. No cloud.

Maziyar PANAHI

329,986 次观看 • 4 个月前

Gemma 4 looks at a parking lot. Decides what to ask. Calls SAM 3.1. "Segment all vehicles." 64 found. "Now just the white ones." 23 found. One model reasoning and orchestrating. One model executing. Both running locally on a MacBook. MLX. No cloud. No API.

Gemma 4 looks at a parking lot. Decides what to ask. Calls SAM 3.1. "Segment all vehicles." 64 found. "Now just the white ones." 23 found. One model reasoning and orchestrating. One model executing. Both running locally on a MacBook. MLX. No cloud. No API.

Maziyar PANAHI

594,212 次观看 • 4 个月前

I finally managed to use Unlimited-OCR and Gemma 4 together in OpenMed. 🔥 A real patient chart: read, de-identified, and mapped to FHIR on one laptop. No cloud, no API key, nothing leaves the machine. All via llama.cpp, all free on Hugging Face. 🤗 What do we do next?

I finally managed to use Unlimited-OCR and Gemma 4 together in OpenMed. 🔥 A real patient chart: read, de-identified, and mapped to FHIR on one laptop. No cloud, no API key, nothing leaves the machine. All via llama.cpp, all free on Hugging Face. 🤗 What do we do next?

Maziyar PANAHI

42,263 次观看 • 1 个月前

Wow! This is amazing! Segmented every car locally in real time with Meta's SAM3 converted to MLX. Just on-device (M2 laptop) vision getting absurdly good. Local AI is moving faster than most people realize! What other models should we test? what kind of videos?

Wow! This is amazing! Segmented every car locally in real time with Meta's SAM3 converted to MLX. Just on-device (M2 laptop) vision getting absurdly good. Local AI is moving faster than most people realize! What other models should we test? what kind of videos?

Maziyar PANAHI

177,270 次观看 • 4 个月前

Claude fable 5 is cooked Claude fable 5 Vs Gemma 4 26b a4b qat MTP Gemma 4 26b running locally on my 8GB vram single RTX 4060 built this using three.js in a single session and 3 prompts. no cloud, no subscription 100% private unlimited use. how long until it can catch up completely?

Claude fable 5 is cooked Claude fable 5 Vs Gemma 4 26b a4b qat MTP Gemma 4 26b running locally on my 8GB vram single RTX 4060 built this using three.js in a single session and 3 prompts. no cloud, no subscription 100% private unlimited use. how long until it can catch up completely?

Alok

30,802 次观看 • 2 个月前

Cancelled ChatGPT -> Built JARVIS -> Pays $0 -> it works offline + it's smarter than the $20/month version. No WiFi needed, no cloud, no API keys, no rate limits, no queues, no $20/month just to ask a server in Virginia for the weather. Just a local model running directly on the laptop hardware, voice activated, system integrated, controlling apps, answering questions, doing the work. Iron Man had JARVIS embedded in his suit, this guy has it embedded in his MacBook and it works on a plane, in a basement, on a remote cabin with zero signal. OpenAI is burning $700,000 a day on infrastructure to deliver something this guy runs for free. Anthropic charges $200/month for unlimited Claude access, microsoft built Copilot into every product they sell. This guy skipped all of it, downloaded a model and made his laptop the smartest device in the room. No subscription. No login. No internet. No data sent anywhere ever. The most powerful AI assistant on earth is now the one running locally on hardware you already own. ChatGPT charges you to think slower, he pays nothing and thinks alone, he made it himself.

Cancelled ChatGPT -> Built JARVIS -> Pays $0 -> it works offline + it's smarter than the $20/month version. No WiFi needed, no cloud, no API keys, no rate limits, no queues, no $20/month just to ask a server in Virginia for the weather. Just a local model running directly on the laptop hardware, voice activated, system integrated, controlling apps, answering questions, doing the work. Iron Man had JARVIS embedded in his suit, this guy has it embedded in his MacBook and it works on a plane, in a basement, on a remote cabin with zero signal. OpenAI is burning $700,000 a day on infrastructure to deliver something this guy runs for free. Anthropic charges $200/month for unlimited Claude access, microsoft built Copilot into every product they sell. This guy skipped all of it, downloaded a model and made his laptop the smartest device in the room. No subscription. No login. No internet. No data sent anywhere ever. The most powerful AI assistant on earth is now the one running locally on hardware you already own. ChatGPT charges you to think slower, he pays nothing and thinks alone, he made it himself.

Defileo🔮

154,465 次观看 • 3 个月前

An unusual object was spotted over Kyiv. Such objects in the sky surprise and concern Ukrainians at the same time It is most likely a comet. No air-raid alert was declared 📹 Sternenko

An unusual object was spotted over Kyiv. Such objects in the sky surprise and concern Ukrainians at the same time It is most likely a comet. No air-raid alert was declared 📹 Sternenko

Euromaidan Press

24,005 次观看 • 4 个月前

GEMMA 4 26B ON AN RTX 4060 WITH A 248K TOKEN CONTEXT WINDOW 20 tokens per second and a context window so large you can feed it entire codebases, books and research papers in a single prompt this is not a cloud api and not a server rack, this is a regular consumer gpu running locally with llama.cpp and q4_k_xl quantization 248k context on an 8gb vram card was not supposed to be possible and here it is just running on someone’s desk the article below covers exactly which tools and configs make this kind of setup work in 2026 ↓

GEMMA 4 26B ON AN RTX 4060 WITH A 248K TOKEN CONTEXT WINDOW 20 tokens per second and a context window so large you can feed it entire codebases, books and research papers in a single prompt this is not a cloud api and not a server rack, this is a regular consumer gpu running locally with llama.cpp and q4_k_xl quantization 248k context on an 8gb vram card was not supposed to be possible and here it is just running on someone’s desk the article below covers exactly which tools and configs make this kind of setup work in 2026 ↓

leopardracer

56,772 次观看 • 2 个月前

I showed you SAM 3 all week. This is a 0.6B model that outperforms it. Falcon Perception. Type "detect the plane" and it segments every plane in the frame. Pixel-accurate masks from natural language. Fighter jets. Fire. Crowds. All on a MacBook via MLX. No cloud.

I showed you SAM 3 all week. This is a 0.6B model that outperforms it. Falcon Perception. Type "detect the plane" and it segments every plane in the frame. Pixel-accurate masks from natural language. Fighter jets. Fire. Crowds. All on a MacBook via MLX. No cloud.

Maziyar PANAHI

63,029 次观看 • 4 个月前

Meet Gemma 3n, a model that runs on as little as 2GB of RAM 🤯 It shares the same architecture as Gemini Nano, and is engineered for incredible performance. We added audio understanding, so now it’s multimodal, fast and lean, and runs on-device (no cloud connection required!)

Meet Gemma 3n, a model that runs on as little as 2GB of RAM 🤯 It shares the same architecture as Gemini Nano, and is engineered for incredible performance. We added audio understanding, so now it’s multimodal, fast and lean, and runs on-device (no cloud connection required!)

Google AI

228,835 次观看 • 1 年前

AN $8 ESP32 IS RUNNING LOCAL AI WITH ZERO LATENCY A creator deployed a computer vision model on an ESP32 microcontroller to track human movement locally in real time, showing what’s possible without a GPU.

AN $8 ESP32 IS RUNNING LOCAL AI WITH ZERO LATENCY A creator deployed a computer vision model on an ESP32 microcontroller to track human movement locally in real time, showing what’s possible without a GPU.

0xMarioNawfal

58,032 次观看 • 26 天前

Kimi K3 beats GPT 5.6 Sol on both speed and in cost in backend bug fixing ! Planted 4 defects in the same Python order service: a wrong-total calculation, a broken access-control check (IDOR), a crash on invalid input, and an inventory oversell. Both models got identical code and the same prompt, with an automated test suite verifying every fix over the live API. Both went 4/4, but by the time Kimi K3 finished all 4, GPT 5.6 had only fixed 2. Kimi wrapped in under half the time, at half the cost. Kimi K3: 1min 34s · $0.1 · 4/4 GPT 5.6 Sol: 4min 6s · $0.2 · 4/4

Kimi K3 beats GPT 5.6 Sol on both speed and in cost in backend bug fixing ! Planted 4 defects in the same Python order service: a wrong-total calculation, a broken access-control check (IDOR), a crash on invalid input, and an inventory oversell. Both models got identical code and the same prompt, with an automated test suite verifying every fix over the live API. Both went 4/4, but by the time Kimi K3 finished all 4, GPT 5.6 had only fixed 2. Kimi wrapped in under half the time, at half the cost. Kimi K3: 1min 34s · $0.1 · 4/4 GPT 5.6 Sol: 4min 6s · $0.2 · 4/4

GMI Cloud

24,606 次观看 • 1 个月前

Every creator has been here: an idea in your head, no time to build it, no clean way to visualize it. So I tried something different… I opened Freepik and built the entire concept as a short video, start to finish, in one place. I wrote a simple scene, dropped in my product, and let Freepik generate a cinematic first draft. Then I refined it: better framing, smoother motion, cleaner lighting. The kind of polish you’d normally need a full editing stack for, all handled inside the same workflow. One idea. One tool. A finished video in minutes. Try it now:

Every creator has been here: an idea in your head, no time to build it, no clean way to visualize it. So I tried something different… I opened Freepik and built the entire concept as a short video, start to finish, in one place. I wrote a simple scene, dropped in my product, and let Freepik generate a cinematic first draft. Then I refined it: better framing, smoother motion, cleaner lighting. The kind of polish you’d normally need a full editing stack for, all handled inside the same workflow. One idea. One tool. A finished video in minutes. Try it now:

Chidanand Tripathi

93,250 次观看 • 9 个月前

You know what time it is. Siraya Tech ODG PET-CF gets the hammer test. Keep in mind if I hit a regular Glock frame this hard, the result would be the same. No. I won’t.

You know what time it is. Siraya Tech ODG PET-CF gets the hammer test. Keep in mind if I hit a regular Glock frame this hard, the result would be the same. No. I won’t.

StayFree3DP

18,579 次观看 • 9 个月前

🤔 Imagine if your backend worked like `useState`. No API calls. No fetching. Just reactive, real-time state that syncs across devices. That’s what InstantDB is doing. With Instant your frontend gets a real-time database. You write queries that stay in sync and work offline. Every save updates immediately, and it comes with built-in auth and permissions. See how InstantDB looks in action in this deeply technical blog post from Stopa ↓

🤔 Imagine if your backend worked like `useState`. No API calls. No fetching. Just reactive, real-time state that syncs across devices. That’s what InstantDB is doing. With Instant your frontend gets a real-time database. You write queries that stay in sync and work offline. Every save updates immediately, and it comes with built-in auth and permissions. See how InstantDB looks in action in this deeply technical blog post from Stopa ↓

Expo

53,481 次观看 • 1 年前

Three boys found gold in Cheporor on Saturday. By Tuesday, 10,000 people were standing in that same village. From all over West Pokot, and from Uganda. I know Kanyarkwat. It is quiet. There is no water for 10,000 people. No toilets. No police post for a crowd that size. The gold is real. The danger is also real.

Three boys found gold in Cheporor on Saturday. By Tuesday, 10,000 people were standing in that same village. From all over West Pokot, and from Uganda. I know Kanyarkwat. It is quiet. There is no water for 10,000 people. No toilets. No police post for a crowd that size. The gold is real. The danger is also real.

Ruto James

103,539 次观看 • 19 天前

AI video has changed so fast most people don't realize it yet The average agency still charges $500 for one UGC ad with a real human, lighting setup, post-production, the works I just made one in under 20 minutes. Character, voice, scene, everything generated. Same output. 1/25 the cost. No casting, no shoots, no studio time. RT this and comment "prompt" and I'll DM you the exact workflow I used.

AI video has changed so fast most people don't realize it yet The average agency still charges $500 for one UGC ad with a real human, lighting setup, post-production, the works I just made one in under 20 minutes. Character, voice, scene, everything generated. Same output. 1/25 the cost. No casting, no shoots, no studio time. RT this and comment "prompt" and I'll DM you the exact workflow I used.

Kiyoro

132,410 次观看 • 3 个月前

Web Scraping is dead. Web Agenting is here. Writing selectors (div > .class > span) breaks every time a site updates. Building custom bots for every new target is a waste of engineering hours. TinyFish turns the entire real-time web into a single API. Input: Natural Language ("Find availability for X"). Target: 1 or 100 URLs. Output: Structured JSON. This isn't a simulation. It visits the Real-Time Web. 1./ One API, Many Sites - Same contract whether you hit 1 URL or 50. You focus on the Goal (Business Logic). TinyFish handles the How (Navigation, Clicks, Inputs). 2./ Real Automation - It doesn't just "read." It interacts. It fills forms, navigates multi-step flows, and handles dynamic JS content. 3./ Production Ready - This is the same infrastructure used by large enterprises, now exposed as a clean developer primitive. Logs, error handling, and structured data are built-in. The web is finally a proper API.

Web Scraping is dead. Web Agenting is here. Writing selectors (div > .class > span) breaks every time a site updates. Building custom bots for every new target is a waste of engineering hours. TinyFish turns the entire real-time web into a single API. Input: Natural Language ("Find availability for X"). Target: 1 or 100 URLs. Output: Structured JSON. This isn't a simulation. It visits the Real-Time Web. 1./ One API, Many Sites - Same contract whether you hit 1 URL or 50. You focus on the Goal (Business Logic). TinyFish handles the How (Navigation, Clicks, Inputs). 2./ Real Automation - It doesn't just "read." It interacts. It fills forms, navigates multi-step flows, and handles dynamic JS content. 3./ Production Ready - This is the same infrastructure used by large enterprises, now exposed as a clean developer primitive. Logs, error handling, and structured data are built-in. The web is finally a proper API.

Tech with Mak

119,743 次观看 • 6 个月前

i just ran Google's brand new Unsloth Gemma4 12B dense GGUF on my RTX 4060 using llama.cpp + CUDA 13.2 21 tokens per second. on a budget consumer GPU. locally. no API. no cloud. no subscription. and the benchmarks are absolutely cooked # first let's talk architecture because this is genuinely different every multimodal model you've used has a frozen vision encoder + frozen audio encoder + LLM backbone glued together Gemma 4 12B is different it's a single decoder only transformer. that's it. vision? raw 48×48 pixel patches → one matmul → projected directly into the LLM audio? raw 16kHz signal sliced into 40ms frames → linear projection → same LLM input space no encoder tax. no latency penalty. no fragmented memory to put the encoder savings in perspective: old Gemma 4 26B approach: - 550M param vision encoder (frozen) - 300M param audio encoder (frozen) - LLM backbone Gemma 4 12B: - 35M param vision embedder (a single matmul) - no audio encoder at all - LLM backbone handles EVERYTHING 550M → 35M for vision alone. that's a 15x reduction this is why the gemma-4-12b-it-Q4_K_M.gguf is just 6.6 GBs!!! and it has 256K native context context # Benchmarks: AIME 2026 (math olympiad): 77.5% GPQA Diamond (expert science): 78.8% LiveCodeBench v6 (real code): 72% Codeforces ELO: 1659 MMLU Pro: 77.2% MATH-Vision: 79.7% BigBench Extra Hard: 53% inference → llama.cpp, LM Studio, vLLM, SGLang llamacpp flags: -m "gemma-4-12b-it-Q4_K_M.gguf" -ngl 99 -c 8000 -v --port 8080 Available on huggingface now! Link below

Alok

280,546 次观看 • 2 个月前