Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

generated multi-modal embeddings of my bookmarks using Google DeepMind’s new Gemini Embedding 2, allowing low-dim spatialisation using UMAP and high-dim similarity querying via different media types.

harley turan

12,956 subscribers

25,291 просмотров • 5 месяцев назад •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

Integrate Gemini and Imagen APIs into your Flutter applications using the `firebase_ai` package 🪄🔥 In this video, you'll discover different model types for generative content and image generation, constructing multi-modal prompts, and handling responses.

Integrate Gemini and Imagen APIs into your Flutter applications using the `firebase_ai` package 🪄🔥 In this video, you'll discover different model types for generative content and image generation, constructing multi-modal prompts, and handling responses.

Flutter

14,874 просмотров • 5 месяцев назад

$We’re shipping two major updates to streamline your creative workflow, allowing you to generate high-speed images with one model and then instantly animate them with the other—all at a fraction of the cost 🍌⚡️ 1️⃣ Introducing Nano Banana 2 Lite: Our fastest and most cost-efficient Gemini Image model yet delivers text-to-image outputs in under 4 seconds. Now available via the Gemini API and Google AI Studio, and rolling out soon across @NotebookLM, Google Flow, Google Gemini, Stitch by Google, Google Search and Google Photos. 2️⃣ Gemini Omni Flash in Public Preview: Our natively multimodal model for cost-efficient video generation and conversational editing. Now available via the Gemini API, Google AI Studio, and Gemini Enterprise Agent Platform so you can integrate the model into your workflow. While exciting on their own, the real magic happens when you build using these models together. Watch how our interior design demo integrates Nano Banana 2 Lite and Omni to instantly reimagine any space. Upload a photo, swipe through tailored design concepts, and see Omni bring the details to life in cinematic motion. Try out the demo app in AI Studio:$

We’re shipping two major updates to streamline your creative workflow, allowing you to generate high-speed images with one model and then instantly animate them with the other—all at a fraction of the cost 🍌⚡️ 1️⃣ Introducing Nano Banana 2 Lite: Our fastest and most cost-efficient Gemini Image model yet delivers text-to-image outputs in under 4 seconds. Now available via the Gemini API and Google AI Studio, and rolling out soon across @NotebookLM, Google Flow, Google Gemini, Stitch by Google, Google Search and Google Photos. 2️⃣ Gemini Omni Flash in Public Preview: Our natively multimodal model for cost-efficient video generation and conversational editing. Now available via the Gemini API, Google AI Studio, and Gemini Enterprise Agent Platform so you can integrate the model into your workflow. While exciting on their own, the real magic happens when you build using these models together. Watch how our interior design demo integrates Nano Banana 2 Lite and Omni to instantly reimagine any space. Upload a photo, swipe through tailored design concepts, and see Omni bring the details to life in cinematic motion. Try out the demo app in AI Studio:

Google AI

129,052 просмотров • 2 месяцев назад

New Example! We built a fullstack open “Deep Research” quickstart for Google DeepMind Gemini 2.5. It dynamically searches the web, reflects on results, and delivers comprehensive answers with citations in a nice UI with streaming! Built using React and LangChain Langgraph! 🚀 TL;DR: 🔄 Agent iteratively loops through research and reflection until it gathers sufficient information. 🔍 Dynamic query generation, web research via Gemini native Google Search tool, and reflective reasoning. 🧠 Supports different search effors (low, medium, high) for width and depth of search 🛠️ React frontend, LangGraph backend, Tailwind CSS + Shadcn UI components. 🐳 Easily run locally or deploy with Docker. 📄 Answers inlcude citations from gathered web sources.

New Example! We built a fullstack open “Deep Research” quickstart for Google DeepMind Gemini 2.5. It dynamically searches the web, reflects on results, and delivers comprehensive answers with citations in a nice UI with streaming! Built using React and LangChain Langgraph! 🚀 TL;DR: 🔄 Agent iteratively loops through research and reflection until it gathers sufficient information. 🔍 Dynamic query generation, web research via Gemini native Google Search tool, and reflective reasoning. 🧠 Supports different search effors (low, medium, high) for width and depth of search 🛠️ React frontend, LangGraph backend, Tailwind CSS + Shadcn UI components. 🐳 Easily run locally or deploy with Docker. 📄 Answers inlcude citations from gathered web sources.

Philipp Schmid

55,535 просмотров • 1 год назад

This is, by far, one of the best uses of modern AI. If you don't use embeddings when querying your database, you are definitely leaving a lot on the table. In this video, I'll show you how to run semantic searches using OpenAI and PostgreSQL. It's all thanks to Pgai, an open-source PostgreSQL extension: Here's what will happen: 1. We'll create a simple table with news articles 2. We'll generate embeddings for those articles 3. We'll run queries on top of those embeddings For this video, I generated the embeddings using a simple query, but pgai Vectorizer would do the same automatically as new information makes it into the database. This is awesome! If you have a PostgreSQL database with data you are searching over, you should start experimenting with semantic searches immediately. For most use cases, a combination of full-text search + semantic search is the best approach. If you don't have a PostgreSQL database around, you can try free for 30 days using Timescale: Thanks to the TimescaleDB (by Tiger Data) team for partnering with me on this post!

This is, by far, one of the best uses of modern AI. If you don't use embeddings when querying your database, you are definitely leaving a lot on the table. In this video, I'll show you how to run semantic searches using OpenAI and PostgreSQL. It's all thanks to Pgai, an open-source PostgreSQL extension: Here's what will happen: 1. We'll create a simple table with news articles 2. We'll generate embeddings for those articles 3. We'll run queries on top of those embeddings For this video, I generated the embeddings using a simple query, but pgai Vectorizer would do the same automatically as new information makes it into the database. This is awesome! If you have a PostgreSQL database with data you are searching over, you should start experimenting with semantic searches immediately. For most use cases, a combination of full-text search + semantic search is the best approach. If you don't have a PostgreSQL database around, you can try free for 30 days using Timescale: Thanks to the TimescaleDB (by Tiger Data) team for partnering with me on this post!

Santiago

110,133 просмотров • 1 год назад

Introducing Nano Banana 2, our best image model yet 🍌🍌 It uses Gemini’s understanding of the world and is powered by real-time information and images from web search. That means it can better reflect real-world conditions in high-fidelity. Check out "Window Seat," a demo using Nano Banana 2’s world understanding to generate more accurate views from any window in the world, pulling live local weather info with 2K/4K specs. The precision is mind blowing. Rolling out today as the new default in the Google Gemini, Search (across 141 countries), and Flow + available in preview via Google AI Studio and Vertex AI. Also available in Google Google Antigravity.

Introducing Nano Banana 2, our best image model yet 🍌🍌 It uses Gemini’s understanding of the world and is powered by real-time information and images from web search. That means it can better reflect real-world conditions in high-fidelity. Check out "Window Seat," a demo using Nano Banana 2’s world understanding to generate more accurate views from any window in the world, pulling live local weather info with 2K/4K specs. The precision is mind blowing. Rolling out today as the new default in the Google Gemini, Search (across 141 countries), and Flow + available in preview via Google AI Studio and Vertex AI. Also available in Google Google Antigravity.

Sundar Pichai

923,505 просмотров • 6 месяцев назад

So far, this is one of my most favorite iOS 27 features - wallpaper extend The generated results are so extremely high quality and real-looking that I even see myself using it. It’s still buggy and works like 1/2 of the time, but it has a HUGE future

So far, this is one of my most favorite iOS 27 features - wallpaper extend The generated results are so extremely high quality and real-looking that I even see myself using it. It’s still buggy and works like 1/2 of the time, but it has a HUGE future

Out of Galaxy

721,085 просмотров • 2 месяцев назад

Sharing this n8n template for $0 (FREE) I automatically track competitors' Facebook ad campaigns daily using this n8n workflow I created. - Gets the latest 10 ads from competitors - Analyzes video visuals and audio transcription using Gemini - Analyzes images using OpenAI (4o-mini) - Saves everything in a Google Sheet I'll be listing this workflow for $25 in the coming days. Exclusively FREE for my people for the next 2 days. Follow + Repost + Reply “ADS” for the n8n workflow via DM (free, no email needed).

Sharing this n8n template for $0 (FREE) I automatically track competitors' Facebook ad campaigns daily using this n8n workflow I created. - Gets the latest 10 ads from competitors - Analyzes video visuals and audio transcription using Gemini - Analyzes images using OpenAI (4o-mini) - Saves everything in a Google Sheet I'll be listing this workflow for $25 in the coming days. Exclusively FREE for my people for the next 2 days. Follow + Repost + Reply “ADS” for the n8n workflow via DM (free, no email needed).

Archit Jain

89,073 просмотров • 1 год назад

Google just dropped its most powerful model and GPT-4's biggest competitor, Gemini Ultra. Here's what stands out: 1. You can opt out of training data when signing up 2. There is a digital watermark applied to all images generated by Imagen 2 within Gemini Ultra (but you can't see it) 3. Ultra is much more capable of complex tasks than Gemini Pro, such as coding, logical reasoning, and following longer/more detailed instructions. 4. There's a brand new Google Gemini App on Google Play and the App Store (in the US only for now) 5. Gemini on Android supports some existing Google Assistant features like controlling your Home devices, setting alarms, and more 6. Gemini on Android can pull data from Google Flights, Google Hotels, Google Maps, Workspace, and YouTube and be invoked by saying “Hey Google” 7. Bard is now fully rebranded as Gemini (RIP Bard) 8. It’s finally coming to Canada 🇨🇦 9. Gemini Ultra is priced at $20/mo (same price as ChatGPT) but has a 2-month free trial I'll be testing it out and sharing more thoughts in my newsletter tomorrow morning. If you've tried it, what are your first impressions? Let me know where it performed better/worse than GPT-4.

Rowan Cheung

160,813 просмотров • 2 лет назад

New short course: Building Multimodal Search and RAG", by Weaviate AI Database's Sebastia(N_) Witalec ✊🏽✊🏾✊🏿. Contrastive learning is used to train models to map vectors into an embedding space by pulling similar concepts closer together and pushing dissimilar concepts away from each other. This technique is also used to train multimodal embedding models that capture semantic similarity across different modalities like text, images, and audio. These multimodal embeddings can be used to build multimodal search and RAG systems. In this course, you'll learn how contrastive learning works, and how to add multimodality to RAG – so your models can draw on diverse, relevant context to answer questions. For example, a query about a financial report might synthesize information from text snippets, graphs, tables, and slides. You will also learn how visual instruction tuning lets you integrate image understanding into language models, and build a multi-vector recommender system using Weaviate’s open-source vector database. Please sign up here:

New short course: Building Multimodal Search and RAG", by Weaviate AI Database's Sebastia(N_) Witalec ✊🏽✊🏾✊🏿. Contrastive learning is used to train models to map vectors into an embedding space by pulling similar concepts closer together and pushing dissimilar concepts away from each other. This technique is also used to train multimodal embedding models that capture semantic similarity across different modalities like text, images, and audio. These multimodal embeddings can be used to build multimodal search and RAG systems. In this course, you'll learn how contrastive learning works, and how to add multimodality to RAG – so your models can draw on diverse, relevant context to answer questions. For example, a query about a financial report might synthesize information from text snippets, graphs, tables, and slides. You will also learn how visual instruction tuning lets you integrate image understanding into language models, and build a multi-vector recommender system using Weaviate’s open-source vector database. Please sign up here:

Andrew Ng

104,371 просмотров • 2 лет назад

NEWS: Waymo has introduced the Waymo World Model, "a frontier generative model built on Google DeepMind’s Genie 3 that sets a new bar for large-scale, hyper-realistic autonomous driving simulation." "By simulating the “impossible”, we proactively prepare the Waymo Driver for some of the most rare and complex scenarios—from tornadoes to planes landing on freeways—long before it encounters them in the real world. The model’s architecture offers high controllability, allowing our engineers to modify simulations with simple language prompts, driving inputs, and scene layouts. Notably, the Waymo World Model generates high-fidelity, multi-sensor outputs that include both camera and lidar data. This combination of broad world knowledge, fine-grained controllability, and multi-modal realism enhances Waymo’s ability to safely scale our service across more places and new driving environments. In the following sections we showcase the Waymo World Model in action, featuring simulations of the Waymo Driver navigating diverse rare edge-case scenarios." More info below.

NEWS: Waymo has introduced the Waymo World Model, "a frontier generative model built on Google DeepMind’s Genie 3 that sets a new bar for large-scale, hyper-realistic autonomous driving simulation." "By simulating the “impossible”, we proactively prepare the Waymo Driver for some of the most rare and complex scenarios—from tornadoes to planes landing on freeways—long before it encounters them in the real world. The model’s architecture offers high controllability, allowing our engineers to modify simulations with simple language prompts, driving inputs, and scene layouts. Notably, the Waymo World Model generates high-fidelity, multi-sensor outputs that include both camera and lidar data. This combination of broad world knowledge, fine-grained controllability, and multi-modal realism enhances Waymo’s ability to safely scale our service across more places and new driving environments. In the following sections we showcase the Waymo World Model in action, featuring simulations of the Waymo Driver navigating diverse rare edge-case scenarios." More info below.

Sawyer Merritt

133,168 просмотров • 6 месяцев назад

Google just proved that bigger isn't always better. Their 308M parameter model is outperforming models 2x its size. Google just released 𝗘𝗺𝗯𝗲𝗱𝗱𝗶𝗻𝗴𝗚𝗲𝗺𝗺𝗮, and it's proving that lightweight embedding models can punch way above their weight class. At just 308M parameters (578MB), it's the new state-of-the-art for models under 500M parameters across MTEB multilingual, English, and code benchmarks. But the really impressive part is that it ranks 8th overall on MTEB(Multilingual, v2) - that's 𝟭𝟳 𝗽𝗹𝗮𝗰𝗲𝘀 above the second-best sub-500M model, and it's delivering performance 𝗰𝗼𝗺𝗽𝗮𝗿𝗮𝗯𝗹𝗲 𝘁𝗼 𝗺𝗼𝗱𝗲𝗹𝘀 𝗻𝗲𝗮𝗿𝗹𝘆 𝗱𝗼𝘂𝗯𝗹𝗲 𝗶𝘁𝘀 𝘀𝗶𝘇𝗲. There are three key parts of their training recipe that sets it apart: 𝟭. 𝗘𝗻𝗰𝗼𝗱𝗲𝗿-𝗗𝗲𝗰𝗼𝗱𝗲𝗿 𝗜𝗻𝗶𝘁𝗶𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 Instead of starting from a decoder-only Gemma 3 model, they first adapted it to encoder-decoder, then used just the encoder. By basing EmbeddingGemma off an LLM that already has world and language understanding, it gives it a stronger starting point. 𝟮. 𝗧𝗵𝗿𝗲𝗲-𝗟𝗼𝘀𝘀 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 They combine three different loss functions, instead of just having one: • Contrastive loss (NCE) with in-batch negatives and hardness weighting • Spread-out regularization to ensure embeddings utilize the full space (for quantization and ANN retrieval) • Embedding matching distillation from Gemini Embedding - not just learning from relevance scores, but directly aligning the embedding space with the teacher model 𝟯. 𝗠𝗼𝗱𝗲𝗹 𝗦𝗼𝘂𝗽𝗶𝗻𝗴 Rather than just averaging checkpoints from the same training run, they use optimization techniques to find multiple specialized training mixtures. Each mixture creates an "expert" model in different domains, and averaging all their parameters creates a final model that's actually better than individual models. Extras: • Matryoshka embeddings supporting 768, 512, 256, and 128 dimensions • Quantization-aware training - maintains quality even at int4 precision • 100+ languages from Gemma 3 pretraining • Exceptional performance on low-resource languages (check their XTREME-UP results) Is it the absolute best embedding model? No - Gemini Embedding still leads overall. But that's not really the point. EmbeddingGemma proves you can achieve state-of-the-art performance in a small package that's actually deployable on-device, in low-latency applications, and in resource-constrained environments. This makes good embeddings accessible for use cases that I'm seeing more and more: offline applications, privacy-sensitive deployments, and high-throughput scenarios where inference cost actually matters. Full paper: Shoutout to the EmbeddingGemma team at Google DeepMind for this awesome open source work 💙 and to Daniel Williams for helping me with this video! 🫶

Google just proved that bigger isn't always better. Their 308M parameter model is outperforming models 2x its size. Google just released 𝗘𝗺𝗯𝗲𝗱𝗱𝗶𝗻𝗴𝗚𝗲𝗺𝗺𝗮, and it's proving that lightweight embedding models can punch way above their weight class. At just 308M parameters (578MB), it's the new state-of-the-art for models under 500M parameters across MTEB multilingual, English, and code benchmarks. But the really impressive part is that it ranks 8th overall on MTEB(Multilingual, v2) - that's 𝟭𝟳 𝗽𝗹𝗮𝗰𝗲𝘀 above the second-best sub-500M model, and it's delivering performance 𝗰𝗼𝗺𝗽𝗮𝗿𝗮𝗯𝗹𝗲 𝘁𝗼 𝗺𝗼𝗱𝗲𝗹𝘀 𝗻𝗲𝗮𝗿𝗹𝘆 𝗱𝗼𝘂𝗯𝗹𝗲 𝗶𝘁𝘀 𝘀𝗶𝘇𝗲. There are three key parts of their training recipe that sets it apart: 𝟭. 𝗘𝗻𝗰𝗼𝗱𝗲𝗿-𝗗𝗲𝗰𝗼𝗱𝗲𝗿 𝗜𝗻𝗶𝘁𝗶𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 Instead of starting from a decoder-only Gemma 3 model, they first adapted it to encoder-decoder, then used just the encoder. By basing EmbeddingGemma off an LLM that already has world and language understanding, it gives it a stronger starting point. 𝟮. 𝗧𝗵𝗿𝗲𝗲-𝗟𝗼𝘀𝘀 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 They combine three different loss functions, instead of just having one: • Contrastive loss (NCE) with in-batch negatives and hardness weighting • Spread-out regularization to ensure embeddings utilize the full space (for quantization and ANN retrieval) • Embedding matching distillation from Gemini Embedding - not just learning from relevance scores, but directly aligning the embedding space with the teacher model 𝟯. 𝗠𝗼𝗱𝗲𝗹 𝗦𝗼𝘂𝗽𝗶𝗻𝗴 Rather than just averaging checkpoints from the same training run, they use optimization techniques to find multiple specialized training mixtures. Each mixture creates an "expert" model in different domains, and averaging all their parameters creates a final model that's actually better than individual models. Extras: • Matryoshka embeddings supporting 768, 512, 256, and 128 dimensions • Quantization-aware training - maintains quality even at int4 precision • 100+ languages from Gemma 3 pretraining • Exceptional performance on low-resource languages (check their XTREME-UP results) Is it the absolute best embedding model? No - Gemini Embedding still leads overall. But that's not really the point. EmbeddingGemma proves you can achieve state-of-the-art performance in a small package that's actually deployable on-device, in low-latency applications, and in resource-constrained environments. This makes good embeddings accessible for use cases that I'm seeing more and more: offline applications, privacy-sensitive deployments, and high-throughput scenarios where inference cost actually matters. Full paper: Shoutout to the EmbeddingGemma team at Google DeepMind for this awesome open source work 💙 and to Daniel Williams for helping me with this video! 🫶

Victoria Slocum

21,610 просмотров • 9 месяцев назад

Sergey has a new habit. He talks to Gemini Live while driving, discussing things like data center power and cost. It’s classic Google dogfooding — obsessively testing your own product. It reminds me of Bill Gates removing his car radio so he could think about Microsoft nonstop. Every founder should have this level of obsession. Sergey also says the Gemini model he’s using in the car is way better than what’s available now. My guess is that Gemini 3 Flash is coming soon!

Sergey has a new habit. He talks to Gemini Live while driving, discussing things like data center power and cost. It’s classic Google dogfooding — obsessively testing your own product. It reminds me of Bill Gates removing his car radio so he could think about Microsoft nonstop. Every founder should have this level of obsession. Sergey also says the Gemini model he’s using in the car is way better than what’s available now. My guess is that Gemini 3 Flash is coming soon!

Yuchen Jin

652,835 просмотров • 8 месяцев назад

Torch memberships are LIVE at What you get: - Exclusive podcasts, programs & TV specials - My daily radio show + the Torch Insider Show instead of commercial breaks - Daily podcasts generated using my 2 proprietary AIs, George AI and Glenn AI - One-on-one conversations with my favorite thinkers Coming soon, you'll have access to: - My ENTIRE catalogue of radio & TV content, going back decades - Glenn AI – answers generated ONLY from what I have said in the past - George AI – answers generated ONLY from one of the largest collections of founding-era documents in the world - Educational curricula generated using George AI, scheduled to launch in 2027 - Original music written and produced by a new team of musicians and songwriters - Insider conventions & meet-ups - Special live events, beginning May 2nd (BIG announcement coming soon) - And much, much more

Torch memberships are LIVE at What you get: - Exclusive podcasts, programs & TV specials - My daily radio show + the Torch Insider Show instead of commercial breaks - Daily podcasts generated using my 2 proprietary AIs, George AI and Glenn AI - One-on-one conversations with my favorite thinkers Coming soon, you'll have access to: - My ENTIRE catalogue of radio & TV content, going back decades - Glenn AI – answers generated ONLY from what I have said in the past - George AI – answers generated ONLY from one of the largest collections of founding-era documents in the world - Educational curricula generated using George AI, scheduled to launch in 2027 - Original music written and produced by a new team of musicians and songwriters - Insider conventions & meet-ups - Special live events, beginning May 2nd (BIG announcement coming soon) - And much, much more

Glenn Beck

32,132 просмотров • 7 месяцев назад

Google reached out to me to create a short film for Google I/O 2025 using Flow, their new AI filmmaking tool. It’s been a magical experience collaborating with one of the most visionary tech teams in the world. Google #GoogleIO The result is Dear “Stranger”, a film that explores the infinite nature of universal love between a grandmother and grandchild across countless parallel worlds. People often ask why I use AI to craft my cinematic visions. For me, it’s not just about cutting-edge tech—it’s about unlocking stories that might otherwise remain untold. AI removes traditional constraints like time, budget, and scale, allowing me to express something deeply personal: the memory of my grandmother. This film is a visual poem—an emotional offering to her. She was the foundation of my dreams, but never had the chance to witness the life I live today. I imagine her pride, had she seen the person I’ve become. Dear “Stranger” captures flickering memories, shared love, and the eternal question: If there are infinite worlds, could we meet again—somewhere, somehow? It is my honor to partner with Google on this film using Flow, built on DeepMind’s most advanced models—Veo, Imagen, and Gemini. Flow enables natural prompting, consistent characters and scenes, and centralized asset management. GoogleLabs #FlowbyGoogle #VEO #EarlyAccess ✨ Try Flow and share the worlds you imagine:

Google reached out to me to create a short film for Google I/O 2025 using Flow, their new AI filmmaking tool. It’s been a magical experience collaborating with one of the most visionary tech teams in the world. Google #GoogleIO The result is Dear “Stranger”, a film that explores the infinite nature of universal love between a grandmother and grandchild across countless parallel worlds. People often ask why I use AI to craft my cinematic visions. For me, it’s not just about cutting-edge tech—it’s about unlocking stories that might otherwise remain untold. AI removes traditional constraints like time, budget, and scale, allowing me to express something deeply personal: the memory of my grandmother. This film is a visual poem—an emotional offering to her. She was the foundation of my dreams, but never had the chance to witness the life I live today. I imagine her pride, had she seen the person I’ve become. Dear “Stranger” captures flickering memories, shared love, and the eternal question: If there are infinite worlds, could we meet again—somewhere, somehow? It is my honor to partner with Google on this film using Flow, built on DeepMind’s most advanced models—Veo, Imagen, and Gemini. Flow enables natural prompting, consistent characters and scenes, and centralized asset management. GoogleLabs #FlowbyGoogle #VEO #EarlyAccess ✨ Try Flow and share the worlds you imagine:

Junie Lau

129,725 просмотров • 1 год назад

Seedance Playblast 2 Render Demo: VFX artist Doug Hogan built a workflow in ComfyUI that takes a raw Playblast and turns it into a finished shot using Seedance 2.0 - no traditional rendering pipeline. → Playblast = motion ground truth → Nanobanana Pro multi-angle grids = asset identity → Gemini node = auto-generated structured prompt → Seedance 2.0 ties it all together Raw CG animation → finished shot in minutes instead of hours. Click the link below for the workflow👇

Seedance Playblast 2 Render Demo: VFX artist Doug Hogan built a workflow in ComfyUI that takes a raw Playblast and turns it into a finished shot using Seedance 2.0 - no traditional rendering pipeline. → Playblast = motion ground truth → Nanobanana Pro multi-angle grids = asset identity → Gemini node = auto-generated structured prompt → Seedance 2.0 ties it all together Raw CG animation → finished shot in minutes instead of hours. Click the link below for the workflow👇

ComfyUI

13,957 просмотров • 2 месяцев назад

ANNOUNCING: Stoked to share my first AI short film 📺 made on my iPhone Using Gemini Omni that was launched earlier this week Check out: “New Kid” It was my childhood dream to be a storyteller and now thx to AI .. anyone can create What stood out to me w/ Omni: > very easy for beginners like me to go from 0 to 1 > respects reference images > character consistency was epic > the story was mine, refined by chatting with Gemini > built this in 5 scenes each of 10 seconds long and yet didn’t spend much money (overall $49 Google AI Pro) Honorary mention: The edit was done on CapCut Google Gemini Google Flow Looking forward to more late night and weekend experiments ❤️

ANNOUNCING: Stoked to share my first AI short film 📺 made on my iPhone Using Gemini Omni that was launched earlier this week Check out: “New Kid” It was my childhood dream to be a storyteller and now thx to AI .. anyone can create What stood out to me w/ Omni: > very easy for beginners like me to go from 0 to 1 > respects reference images > character consistency was epic > the story was mine, refined by chatting with Gemini > built this in 5 scenes each of 10 seconds long and yet didn’t spend much money (overall $49 Google AI Pro) Honorary mention: The edit was done on CapCut Google Gemini Google Flow Looking forward to more late night and weekend experiments ❤️

KP

36,149 просмотров • 3 месяцев назад

curl This is our Meta-Prompt. It allows LLMs to understand our Reader, Embeddings, Reranker, and Classifier APIs for improved codegen. Using the meta-prompt is straightforward. Just copy the prompt into your preferred LLM interface like ChatGPT, Claude, or whatever works for you, add your instructions, and you're set. In this example, we copied the entire prompt into Anthropic Claude and asked it to grab every sentence from Hacker News front page and visualize them using UMAP with matplotlib. This task is nontrivial as it combines multiple APIs from our Search Foundation, like Reader and Embedding where Claude may not have knowledge of. So if you asked Claude directly, it probably wouldn't give an optimal answer. But with the meta-prompt, Claude now has good knowledge about our APIs and can generate much better code! We can copy paste the code directly to Google Colab and with minimum modification, the code just works!

curl This is our Meta-Prompt. It allows LLMs to understand our Reader, Embeddings, Reranker, and Classifier APIs for improved codegen. Using the meta-prompt is straightforward. Just copy the prompt into your preferred LLM interface like ChatGPT, Claude, or whatever works for you, add your instructions, and you're set. In this example, we copied the entire prompt into Anthropic Claude and asked it to grab every sentence from Hacker News front page and visualize them using UMAP with matplotlib. This task is nontrivial as it combines multiple APIs from our Search Foundation, like Reader and Embedding where Claude may not have knowledge of. So if you asked Claude directly, it probably wouldn't give an optimal answer. But with the meta-prompt, Claude now has good knowledge about our APIs and can generate much better code! We can copy paste the code directly to Google Colab and with minimum modification, the code just works!

Jina AI

209,142 просмотров • 1 год назад

New course: Build and Train an LLM with JAX, built in partnership with Google and taught by Chris Achard. JAX is the open-source library behind Google's Gemini, Veo, and other advanced models. This short course teaches you to build and train a 20-million parameter language model from scratch using JAX and its ecosystem of tools. You'll implement a complete MiniGPT-style architecture from scratch, train it, and chat with your finished model through a graphical interface. Skills you'll gain: - Learn JAX's core primitives: automatic differentiation, JIT compilation, and vectorized execution - Build a MiniGPT-style LLM using Flax/NNX, implementing embedding and transformer blocks - Load a pretrained MiniGPT model and run inference through a chat interface Come learn this important software layer for building LLMs!

New course: Build and Train an LLM with JAX, built in partnership with Google and taught by Chris Achard. JAX is the open-source library behind Google's Gemini, Veo, and other advanced models. This short course teaches you to build and train a 20-million parameter language model from scratch using JAX and its ecosystem of tools. You'll implement a complete MiniGPT-style architecture from scratch, train it, and chat with your finished model through a graphical interface. Skills you'll gain: - Learn JAX's core primitives: automatic differentiation, JIT compilation, and vectorized execution - Build a MiniGPT-style LLM using Flax/NNX, implementing embedding and transformer blocks - Load a pretrained MiniGPT model and run inference through a chat interface Come learn this important software layer for building LLMs!

Andrew Ng

193,825 просмотров • 6 месяцев назад

Our very own Alexander Chen created "Hello, world" a short film inspired by the technology of his childhood with Google Labs Flow. Here's a closer look at how he brought his vision to life: 1) All of the clips were generated in Flow using text-to-image prompts except he used image-to-video prompts for the close-up shots to ensure consistency and accuracy (and the images used were generated by Imagen). 2) To evoke the exact feelings of his childhood home, he used Google Street View for visual references then asked the Google Gemini to write descriptions and prompts, then uploaded those into Flow which is what produced such vivid and accurate renderings in the film. 3) Every sound used in the video is generated by Veo 3. However, to blend things together, he used a separate video editing software to extend the head and tail of each clip over the proceeding and following clip, occasionally crossfading the audio to help with transitions.

Our very own Alexander Chen created "Hello, world" a short film inspired by the technology of his childhood with Google Labs Flow. Here's a closer look at how he brought his vision to life: 1) All of the clips were generated in Flow using text-to-image prompts except he used image-to-video prompts for the close-up shots to ensure consistency and accuracy (and the images used were generated by Imagen). 2) To evoke the exact feelings of his childhood home, he used Google Street View for visual references then asked the Google Gemini to write descriptions and prompts, then uploaded those into Flow which is what produced such vivid and accurate renderings in the film. 3) Every sound used in the video is generated by Veo 3. However, to blend things together, he used a separate video editing software to extend the head and tail of each clip over the proceeding and following clip, occasionally crossfading the audio to help with transitions.

Google AI

94,790 просмотров • 1 год назад

People don't realize that factories in Asia and factories in the USA are totally different. I think AI is going to help the Asian factories much more than the American ones. Factories in Asia are high mix. Factories in the USA are low mix. This means that factories in the USA make significantly fewer products, whereas factories in Asia are way more flexible, mainly because of the huge differential in labor costs. The gigafactory in Austin, one of the biggest factories in the world, makes 2 products: Tesla Model Y The Cybertruck That's it. The factory in the video makes 1000s of different products every year. When you're making the same thing over and over again, maybe in a few different colors, you're much better off using normal automation which is lower error than AI and human design. But when you're making 1000s of different products, using AI to spin up product 1001 starts to kind of make sense. There is not time or logic to using traditional automation; you're better off using AI to make a fast quality control mechanism and then repair errors using your inexpensive human labor. People are viewing AI in factories like there is one type of factory, but the way factories run in high cost countries vs. low cost countries is wildly different. We're just at the beginning of what will be an epic change for not only manufacturing, but the cost and quality of the goods we consume. The real internet of things has arrived.

People don't realize that factories in Asia and factories in the USA are totally different. I think AI is going to help the Asian factories much more than the American ones. Factories in Asia are high mix. Factories in the USA are low mix. This means that factories in the USA make significantly fewer products, whereas factories in Asia are way more flexible, mainly because of the huge differential in labor costs. The gigafactory in Austin, one of the biggest factories in the world, makes 2 products: Tesla Model Y The Cybertruck That's it. The factory in the video makes 1000s of different products every year. When you're making the same thing over and over again, maybe in a few different colors, you're much better off using normal automation which is lower error than AI and human design. But when you're making 1000s of different products, using AI to spin up product 1001 starts to kind of make sense. There is not time or logic to using traditional automation; you're better off using AI to make a fast quality control mechanism and then repair errors using your inexpensive human labor. People are viewing AI in factories like there is one type of factory, but the way factories run in high cost countries vs. low cost countries is wildly different. We're just at the beginning of what will be an epic change for not only manufacturing, but the cost and quality of the goods we consume. The real internet of things has arrived.

molson 🧠⚙️

66,129 просмотров • 8 месяцев назад