🚀 Excited to launch HPU training in partnership with Habana Labs! Every ⚡New Handle: @LightningAI ⚡ model already supports it! Trainer(accelerator=“hpu”) 🤯 🤯

A big shout out to everyone contributing to the development of PyTorch. PyTorch PyTorch Live ⚡New Handle: @LightningAI ⚡ Ryan O'Connor I borrowed from your article. assemblyai.com/blog/pytorch-v…

Amazing PyTorch-lightning tutorials with code by Phillip Lippe. Tutorials cover many deep learning topics like transformers, energy-based models, GNNs, and more! We recommend the tutorials for everyone starting their DL journey: pytorch-lightning.readthedocs.io/en/latest/

I've created a template repo for easily using LIGHTNING.AI and Panel to build, scale and deploy powerful #machinelearning #dataapp in #python Check it out github.com/marcskovmadsen… #TheToolsYouKnowAndLove ❤️ ⚡New Handle: @LightningAI ⚡ PyTorch #dataviz #DataScience

Would you like to be able to add a Panel #dataapp to your ⚡New Handle: @LightningAI ⚡ App? Check out my Live App: …xzp3npc9cr2dmthh.litng-ai-03.litng.ai/view/Home Repo: github.com/marcskovmadsen… PR to Lightning Docs: github.com/Lightning-AI/l… PyTorch #python #machinelearning #deeplearning

With just a few days of work, you too can build an intelligent app that is: 📈 Scalable 💪 Performant & distributed 🎨 Fully customizable Learn how 👉👉👉 bit.ly/3fYUf6n #Diffusion #AI #ML #BuildWithLightning

What if you could launch new products and AI startups in only a few days with Lightning AI ⚡️ and our new apps framework! This Lightning app for stable diffusion took 2 weeks to build with 2 engs. Shout out to Emad and Stability AI for providing awesome opensource models

Introducing Boltus: The God of AI ⚡ Binge 🍿 all four episodes in this 🧵! Let Boltus teach you how to deploy diffusion models at scale 👉👉 bit.ly/3YvdWE2 (1/4)

Progress update!🦙🔥🤓 Lit-LLaMA now implements the LLaMA-Adapter method for efficient fine-tuning 🔧⚡️ The core idea can be implemented in about 11 lines of code🤯 (see screenshot) Link to repo👉 github.com/Lightning-AI/l… Link to Adapter paper👉arxiv.org/abs/2303.16199

Announcing FlashAttention-2! We released FlashAttention a year ago, making attn 2-4 faster and is now widely used in most LLM libraries. Recently I’ve been working on the next version: 2x faster than v1, 5-9x vs standard attn, reaching 225 TFLOPs/s training speed on A100. 1/

Learn how to scrape web data to finetune LLMs using Lightning AI Studio. Everything is included, the code, the data, the Python dependencies, etc.. You can start in seconds instead of hours. lightning.ai/lightning-ai/s…

Prepare a 1 trillion token dataset to train LLMs from scratch in under 4 hours instead of days with Lightning AI ⚡️ Studio! Everything is included, the final datasets, the code, dependencies, etc... Get started in seconds as no setup is needed. lightning.ai/lightning-ai/s…

A while ago I complained here about persistent storage in Google Colab. Have been using Lightning AI ⚡️ Studios for a while now for: - Full VSCode (incl. GH Copilot) - Persisted files shared across notebooks - Multi-GPU/node (!!) It's been great. Feels like a remote ML workstation

🚀 Introducing NSA: A Hardware-Aligned and Natively Trainable Sparse Attention mechanism for ultra-fast long-context training & inference! Core components of NSA: • Dynamic hierarchical sparse strategy • Coarse-grained token compression • Fine-grained token selection 💡 With

Here I show you how to finetune and deploy DeepSeek R1 (8B) for < $1.00 in 8 minutes using the AI Hub from Lightning AI ⚡️ ⚡️⚡️

Introducing DeepSeek-R1 optimizations for Blackwell, delivering 25x more revenue at 20x lower cost per token, compared with NVIDIA H100 just four weeks ago. Fueled by TensorRT DeepSeek optimizations for our Blackwell architecture, including FP4 performance with state-of-the-art

How can we use small LLMs to shift more AI workloads onto our laptops and phones? In our paper and open-source code, we pair on-device LLMs (ollama) with frontier LLMs in the cloud (@openai, @together), to solve token-intensive workloads on your 💻 at 17.5% of the cloud cost