Gray Matter
19 posts

Gray Matter
@GrayMatter_AI
We build software companies! 🤖 AI at your service!
Worldwide Katılım Ağustos 2023
7 Takip Edilen9 Takipçiler

This one’s for the Lord of the Rings fans. Sound on subtitles available and, as always, everything you see and hear is AI generated… #LOTR @midjourney @pika_labs
English

Pop Quiz: If Van Gogh had a Twitter, what would his handle be? Best answers get a virtual high five! 🖐️ #FridayFunnies #nftcommunity
English

Introducing CHITA, an optimization-based approach for pruning pre-trained neural networks at scale. Learn how it leverages advances from several fields and outperforms state-of-the-art pruning methods in terms of scalability and performance tradeoffs → goo.gle/3E12kAc
GIF
English

@Ai_Roundup @GoogleAI It is always good to check the newest trends for our AI products!
English

@GrayMatter_AI @GoogleAI What are you looking forward to doing with it the most?
English

Playing around with the 'blueprint drawing' token in Midjourney generates some pretty stunning artwork.
H/T to @AiProGirl for putting it on our radar in her challenge the other day!
Simple prompt share for the two below: quantum computer, blueprint drawing, batik


English

I got so many bosnian gifts from @symphony_is that they don't fit in the photo, you can't even see the baklavas and the t-shirt I'm wearing 🎁😁

English

Here's a small FastAPI release: 0.101.1 🔖
(probably the last version to have only ones and zeros for a while 😅)
🐛 Fix server error logs with invalid data: show the necessary information to understand and fix invalid data. 🤓
fastapi.tiangolo.com/release-notes/
English

We made @FastAPI even faster. ⚡
Meet the FastAPI Agent: It helps you to manage and create endpoints, write tests, and optimize performance.
🔗 Watch it in action here: youtu.be/pyubW0WxwVg

YouTube
English

📰 Checkout The AI Letters, where Fine was featured this week!
letters.graymatter.ai/p/apple-develo…
English

Let’s break that down in simple language:
1. **Bookshelf Analogy**: Imagine a huge bookshelf filled with books. Some books you use all the time, while others you rarely touch. A neural network is like this shelf. Each book represents a "connection" in the network. Over time, we realize some books (connections) aren't that important.
2. **Spring Cleaning**: To make the bookshelf easier to manage, we want to remove the less important books but without losing too much valuable information. This is similar to "pruning" in neural networks – we remove some connections to save space and make the network run faster, especially on small devices like smartphones.
3. **Choosing Books to Remove**: How do we decide which books to remove? One method is simply looking at the book size; smaller ones might seem less important ("magnitude pruning"). Another method is seeing how removing a book affects your overall knowledge ("optimization-based pruning").
4. **CHITA**: Now, there's a super-efficient way called "CHITA" to decide which books to remove. Instead of just looking at each book individually, CHITA thinks about how each book relates to the others. It's like considering both a book's individual importance and its value in a series.
5. **Results**: Using CHITA, the bookshelf still has the most crucial information but takes up way less space. And, in the neural network world, this means our "pruned" network runs faster and still performs really well.
So, CHITA is like a genius librarian who knows exactly which books to keep and which ones to remove, making our bookshelf (or neural network) super efficient.
English

OpenAI has acquired the team at Global Illumination. The entire team has joined OpenAI to work on our core products including ChatGPT. The team previously designed and built products early on at Instagram and Facebook and have also made significant contributions at YouTube, Google, Pixar, Riot Games, and other notable companies. We’re very excited for the impact they’ll have here at OpenAI. openai.com/blog/openai-ac…
English

@OpenAI Awesome. Congrats! Im rooting for our dear ChatGPT to have the power it used to have!
English


