Yep but there's 70B too, issue is how do you get that 40GB file to the user
LLMS
-
Running Llama 3 Locally in Browser with WebGPU GPU
By
–
I got Llama 3 running in my browser using only my GPU with my Wi-Fi switched OFF completely client-side
— @levelsio (@levelsio) 11 mai 2024
WebGPU is a new feature in browsers where JS can use the GPU of the device and apparently you can run LLMs on it too, and it's fast!
My end goal for 🧠… pic.twitter.com/iz1oxeUQNpI got Llama 3 running in my browser using only my GPU with my Wi-Fi switched OFF completely client-side WebGPU is a new feature in browsers where JS can use the GPU of the device and apparently you can run LLMs on it too, and it's fast! My end goal for
-
Two Types of AI Users: Deterministic vs Creative Temperature Preferences
By
–
in short, there's a kind of people that get a general intelligence capable of anything and always set temperature = 0 and then there's another kind of people that are disappointed that temperature maxes out at 2
-
Two Types of AI Users: Deterministic vs Creative Temperature Preferences
By
–
in short, there's a kind of people that get a general intelligence capable of anything and always set temperature = 0 and then there's another kind of people that are disappointed that temperature maxes out at 2
-
AI Model Specs and Instruction Hierarchy Training Framework Released
By
–
yess already dropping model spec and instruction hierarchy alpha “we use this to train our human trainers” lol who has the constitution for ai now
-
Emotional AI Classification Challenges Vocabulary Limitations
By
–
I should also mention that we're big fans of your work – we've simply hit the limitation that sad (wailing grief), sad (depression), sad (holding back tears) are difficult to disentangle if we don't have the full vocabulary/labels.
-
Testing scaling hypothesis with internet-scale data and distribution shift
By
–
If we had billions of dollars to spend we could test this hypothesis empirically — try to fit a big curve on *all the data* — literally everything that's on the Internet, and then test if it can adapt to things that are slightly out of distribution, like ARC tasks. Oh, wait a
-
Free ChatLLM Trial Available Now on Abacus AI
By
–
First month is free, try it out now: http://
chatllm.abacus.ai -
Access Multiple LLMs and Build AI Agents for $10
By
–
Stop paying $50+ every month for ChatGPT, Claude and Gemini.
— Shubham Saboo (@Saboo_Shubham_) 11 mai 2024
With just $10 you can access GPT-4, Claude, Llama 3, Gemini and more LLMs in single AI playground.
Plus, build custom AI agents with RAG to chat with your internal data without writing a single line of Python Code. pic.twitter.com/mFkVuJN3g4Stop paying $50+ every month for ChatGPT, Claude and Gemini. With just $10 you can access GPT-4, Claude, Llama 3, Gemini and more LLMs in single AI playground. Plus, build custom AI agents with RAG to chat with your internal data without writing a single line of Python Code.
-
Sam Altman addresses ChatGPT model performance and accuracy concerns
By
–
Addressing feedback about ChatGPT’s perceived ‘becoming dumber , Sam Altman acknowledges past issues, but assures significant improvements in model performance and accuracy.