Several tech companies have recently announced the release of Thunder, a new compiler for PyTorch. Thunder promises significant speed improvements in training tasks compared to regular PyTorch. It combines various hardware executors to enhance performance and has garnered excitement within the tech community.
Thunder is the number 5 trending repo on Github! 🤯🤯 Thunder speeds up PyTorch models up to 40% (it's early still 😅😅). Contribute and support it so we can make PyTorch models even faster! https://t.co/o0gtM2zaXc https://t.co/9XlpqSZ9iR
What an incredible week! I’m a bit fried but totally worth it 😄 The biggest highlight was shipping Thunder, our new compiler for PyTorch https://t.co/v9gQRo9w4M It’s going to be fun to see it speeding up our customers’ models 🚀 and see what the community builds with it ⚡️
learned about @LightningAI Thunder compiler from @lantiga at @NVIDIAGTC. Thunder is a source-to-source compiler for @PyTorch. It makes PyTorch programs faster by combining and using different hardware executors at once (ie: nvFuser, torch.compile, cuDNN, and TransformerEngine… https://t.co/HApIH4KTVC
Yesterday we released Lightning Thunder, our new source-to-source compiler for @PyTorch. If you love Python and PyTorch as much as I do and like things that help models to run fast, check out our Zero to Thunder tutorial: https://t.co/aTFjIU62px
🚀 We're super excited to announce our Latest Launch 🎉 The new and improved Deep Learning with PyTorch Course (Version 2.0), is available at a whopping 50% OFF for the next 36 hours! https://t.co/6Lp7CjtIVr Don't let this opportunity slip by. Get this amazing DEAL and elevate… https://t.co/K0cOLAGjGY
We just open-sourced Thunder, a new compiler for PyTorch! In LLM training tasks (e.g., Llama 2 7B), it can achieve a 40% speedup compared to regular PyTorch: https://t.co/P6pvGJBugB What's particularly nice is that you can use it *with* (as opposed to *instead of*)… https://t.co/QxtiCb59NK