Key points
- TorchNeuron enters public beta on October 20 under the Apache 2.0 license.
- The tool supports Amazon's Trainium2 and Trainium3 chips with minimal configuration changes.
- Some PyTorch features are not yet supported, and thousands of tests are skipped.
Amazon has launched TorchNeuron, an open-source tool that allows developers to run existing PyTorch code on Amazon's Trainium AI chips without rewriting their software. This launch aims to simplify the process of moving artificial intelligence workloads from Nvidia hardware to Amazon's own custom silicon.
How it unfolded
- Amazon announced TorchNeuron on Monday, stating it would go into public beta on October 20 at the PyTorch Conference in San Jose.
- The company released the framework under the Apache 2.0 license, supporting both Trainium2 and Trainium3 chips.
- Amazon demonstrated that training scripts run at the same rate on Trainium as on standard chips, with only a single device setting change required.
Who is affected
Developers using PyTorch, the most widely used framework for building AI models, can now target Amazon's hardware more easily. Early adopters include voice AI startups Cartesia and Deepgram, as well as the Lieber Institute for Brain Development, which uses the tool for drug research. Hugging Face has added Trainium support to its AI tools, and Uber recently signed an AWS expansion deal to use the chips.
What happens next
The first release of TorchNeuron has limitations, including a lack of support for dynamic input shapes. According to a PyTorch blog post, approximately 22,000 PyTorch tests are skipped in this initial version. Amazon continues its push to compete with Nvidia, whose chips and CUDA software dominate the AI computing market.
The background in plain language
Moving AI code between different hardware chips
What to do and how to stay safe
- Full-disk encryption protects data if the physical device is lost or stolen.
- Mobile device management allows remote wiping without erasing personal user data.
- Separating corporate apps from personal storage reduces liability and privacy concerns.
Mobile devices bypass traditional network defenses, making device-level controls like encryption and management essential for small businesses. Implement full-disk encryption and remote wipe capabilities immediately, and consider delegating ongoing management to a qualified IT provider.
Step-by-step guide: Mobile Device Security for Small Business: Control Without Cost
General security guidance from the Circuit Observer newsroom. It is not confirmed advice from the organisations named in this story.
