Now, we’re excited to share it openly:.📄 Paper: 🤗 Model: 💻 Code: NASA Blog: IBM Blog:

And this was never a solo effort. It was possible only because of collaboration across heliophysicists, engineers & AI researchers. Thank you to all contributors & co-authors 🙏.Your work shaped every stage of SuryaFM. @NASA @IBMResearch @nvidia @huggingface.

Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.

On my side, I focused heavily on engineering the pipeline.• Handling Data, visualisation and Preprocessing.• Fault-tolerant training.• Profiling and Optimising GPU utilisation.• Scaling to 100+ GPUs with Distributed Training.• Optimised dataloaders for 4K multi-channel input.

With parameter-efficient fine-tuning (LoRA), Surya can:.☀️ Forecast solar flares.💨 Predict solar wind.🌌 Segment active regions.📊 Model EUV spectra.And yes — it can even make a visual prediction of a flare event ⚡.

🔍 Generalizable representation: Even though only half the Sun is visible, temporal prediction forces Surya to learn causal patterns → enabling robust downstream tasks.

Solutions:.🖼️ Vision Transformer: Handles large 4K solar images, capturing both spatial structures and global context. ⏳ Temporal pretext task: Instead of static classification, Surya is trained to predict future solar images → the model must learn dynamics.

🌞 High res: 4096 ×4096 pixels at multiple wavelengths. ⚡ Dynamics: It's easy to capture trivial dynamics like rotation, but it's not interesting. 🌋 Rarity of events: Phenomena like solar flares are rare, making them harder to model. 🌓 Limited view: We always see half the Sun!.

🌞 We’re thrilled to announce Surya, the first heliophysics foundation model, trained at native 4K resolution on NASA’s Solar Dynamics Observatory (SDO) data. 300+ TB of multi-channel, multi-modal solar data → one foundation model.

RT @AymericRoucher: > The first ever Foundation weather model: Prithvi WxC enables life-saving weather predictions! 🌍. Hurricane Katrina ki….

RT @npx_prith: As much as I love seeing the progress in AI, I never thought I’d see an AI model with my name 🌎 Thanks NASA for putting Prit….

Excited to see @ClementDelangue highlighting our work! 🌍 Weather and climate AI like Prithvi WxC open up new frontiers for science, and it’s amazing to be part of this shift beyond just LLMs! Thanks for the shoutout!.

RT @SenBillNelson: Today, @NASA and @IBMResearch released a new weather climate model powered by AI to help us better predict severe weathe….

RT @fdaudens: This is really cool. Check out this comparison of the Ida hurricane between ground truth and the AI model's prediction.#AIWea….

RT @BrigitteTousi: Big news! @IBMResearch and @NASAEarth just released a foundation model for weather and climate, Prithvi WxC, on @hugging….

RT @_akhaliq: NASA and IBM present Prithvi WxC. Foundation Model for Weather and Climate. discuss: Triggered by th….

Do checkout or Prithvi WxC 2.3B weather and climate foundation model!. #NASA #IBM #AIforScience #Foundation #model #open #source.

Dive into more details in our WACV'24 paper 📃.Paper: Code (to be released): Congrats to all the co-authors Arindam Dutta, @dripta_rayc, Calvin-Khang Ta, and Prof. Amit Roy for this work. Special Thanks to @yashgarg10226. (5/5)

Our experiments show POISE's superiority 💪in handling occlusions for applications like gait recognition! (4/5).

Why is POISE interesting? It's self-supervised! 🙌 This means it learns to improve without needing costly annotations, making it scalable and practical. And yes, the algorithm is simple and intuitive! (3/5)

Ever wondered how AI can extract human silhouettes even when a person is partially hidden? 🤔 POISE uses a fusion of segmentation models and 2D pose estimation to do just that! (2/5)