Don't miss Luca Antiga ⚡️ of Lightning AI ⚡️ on stage at #PyTorchConf 2024! Explore the list of keynote speakers: hubs.la/Q02Nfys-0. Register & join us September 18-19 in San Francisco! hubs.la/Q02NfyjH0

LitServe offers full control AND fully customizable scaling. Get complete control of every stage of your model serving process—decoding requests, making model predictions, formatting responses, etc. Check out these and other features: github.com/Lightning-AI/L…

I'm in Mexico City this week and we're bringing together the latino AI community (This Thursday 12 sept!) Share with your friends in CDMX! lu.ma/LAI-takesover-…

L40 GPUs just landed on Lightning Studios! ⚡️ ✅ As low as $1.20/hr (1 GPU) ✅ Ideal for inference, finetuning up to 70B params ✅ Great availability, on-demand, interruptible ✅ 48GB VRAM—2x than A10G + L4! ✅ 1, 4 + 8 GPU instances Try them for free! lightning.ai

🤯 4.7x faster and 3x less GPU memory 🚀 Model memory and speed optimization can boost system throughput and save costs 💰 This article explains precision and quantization methods to optimize model servers for cost-effective scalability ➡️ lightning.ai/docs/litserve/…

Thunder Sessions Livestream returns this Friday at 11am EST! Join Luca Antiga ⚡️ and Thomas Viehmann and see how our source to source compiler boosts AI model speeds by 40% 🚀 Catch up on the previous sessions: youtube.com/playlist?list=… And join live session #8 here:

TorchGeo—the top library to analyze satellite data for farming, climate, disasters, and urban planning—now supports fully reproducible tutorials powered by Lightning Studios: ✅ Environment already set up ✅ Getting started notebooks Try TorchGeo today: torchgeo.readthedocs.io/en/stable/tuto…

Build systems that mix models, databases, and data with LitServe, our lightning-fast serving engine—from simple single-model setups to complex multi-modal systems with multiple models. This example shows a compound system that uses a text model to annotate an image model. Try

Experience the power of Pixtral 12B, served seamlessly via LitServe on an L40 GPU. Ready to give it a try?

LitServe enables streaming—when the server returns results incrementally rather than all at once. This let's users to consume server responses while they're still being delivered (like ChatGPT). Here's a quick Llama 3 example: Try it out yourself ➡️ github.com/Lightning-AI/L…

HackMIT is a weekend-long event where students from around the globe work together to design and build innovative technology projects. ✨ Spotted LitServe being used to serve a multimodal model (ControlNet) 💜 A project by Caleb Felix Nguyen Maxi Attiogbe 👏

We’ll see you this week at #PyTorchConf in SF! Mark your calendar to catch our CTO and PyTorch board member, Luca Antiga ⚡️ give a keynote on the Lightning AI open source stack for accelerating the AI lifecycle on Wed @ 1:15 PT. And stop by our booth for demos and giveaways ⚡️

Humbled at how well LitServe has been received ☺️. LitServe augments FastAPI to dramatically simplify serving for AI models with batching, streaming, GPU auto-detection and more! Try it now github.com/Lightning-AI/L…

LitServe is like FastAPI for deep learning models—handles batching, streaming, multiprocessing, and more. 🤯 Deploy language, audio, CV, and multi-modal models in about 10 lines of code. Simple and scalable. Try it out: github.com/Lightning-AI/L…

🚀 Trying out Pixtral - a multimodal LLM from Mistral AI lightning.ai/lightning-ai/s…

See you at #PyTorchConf — Luca Antiga ⚡️ and Sebastian Raschka will be giving talks! 🍻 Join us as we host the official Welcome Reception 📣 Luca Antiga ⚡️ Wed, Sep 18 @ 1:15p | Thu, Sep 19 @ 11:50a 📣 Sebastian Raschka Thu, Sep 19 @ 9:24a ⚡️ Visit us at booth D2 More info ➡️ lightning.ai/events/confere…

The easiest way to deploy a private multi-modal API with the new Pixtral model by Mistral AI in <30 lines of code 🤯 Take it for a spin ➡ lightning.ai/lightning-ai/s…

Your friendly neighbourhood PyTorch Trainer - Lightning AI ⚡️ is officially most used as per the Python Developer Survey ⚡️🔥

Right now Sebastian Raschka of Lightning AI ⚡️ keynotes on navigating the architectural timeline of LLMs #PyTorchConf

No Thunder Sessions livestream this week as we wrap up #PyTorchConf ⚡ Until we’re back next Friday, learn how Luca Antiga ⚡️ and Thomas Viehmann write custom gradient transforms using Lightning Thunder. Catch up on this and previous sessions: youtube.com/playlist?list=…