Comput3 AI 🌋
@comput3ai
The world's fastest GPUs, at the world's lowest prices for AI Training, Model Validation and Inference. CA: J3NrhzUeKBSA3tJQjNq77zqpWJNz3FS9TrX7H7SLKcom
You can use the inference API and our instances (use our hosted load balancer to keep the same base URL for every instance) with a self hosted Open Webui already. You can run it on your local machine in docker. Quite a few people do this already. comput3.gitbook.io/comput3/comput…
tbh thats exactly what folks need. nxt step is hosting your own open webui and plugging the comput3 api in there but for 90% of the users what you're describing is perfect 🤝✨️
Happy rigging! Unirig is live, just fire up a Media Fast workload and you can use it. Only on 👉 launch.comput3.ai Follow the Infra.
Added rembg and unirig to launch.comput3.ai, should be live within the hour.
We'll add the possibility of downloading and booting any GGUF models on our clusters using this container. Stake to get 100h to 1000h GPU hours every month. Run these models privately on your own GPU. Just 100k-1M $COM. Don't try this at home (you dont have an H200 at home). 🌋
@Fitzydke you need to be all over this. We have 100 hours of GPU usage as part of our 100k $COM stake.
Another day, another contribution to Open Source AI in our mission to get to Open Source AGI. This will run on your Windows gaming Laptop, all the way up to H200 clusters. Follow the infra🌋
To speed up running and downloading gigantic models on gigantic gpus, here's a Dockerfile the llama.cpp as a base, uses huggingface_hub's multi threaded downloader to download the models and then run them. Also included a traefik docker-compose (pro tip) github.com/comput3ai/c3-l…
🌋🌋🌋 fun.
the next generation of autofun is finally here live with custom bonding curves, real sniper mitigation, and the option to choose @MeteoraAG bonding time to build your cult 🧵
We are in fact working on an interface that will store conversations ... but it'll store it to your browser's local storage and not on our backend. Clear your browsing history - that data is gone. We're not monetizing users to train new models. We're on the same team🤝
This basically means that OpenAI is saving all your conversations. Open your eyes 👁️ Use @AskVenice or @comput3ai and use local software to store your conversation. You want to own your data and don't wanna fuel companies that monetize the dialogues you have with AI its crucial
Not sure why X streams always stop for us around 50 mins. Youtube stayed up like a champ. And it's higher resolution.
Youtube 1440p: 1440p stream: youtube.com/live/Ld93Z4K7j…