Hugging Face releases TorchTPU
The story so far
Hugging Face has released TorchTPU, making TPU a native backend device for PyTorch. Users can simply select a model from the Hugging Face Hub, specify device_map="tpu", and Transformers, Diffusers, and Kernels can run directly on TPU. TorchTPU adds TPU as a new backend for PyTorch, allowing users to write regular PyTorch code and install it on TPU virtual machines via the command pip install torch-tpu --extra-index-url https://download.pytorch.org/whl/cpu. This brings native TPU support to the Hugging Face ecosystem.
AI-written from the reports · Updated 27 days ago
Coverage timeline
Follow the coverage to see every side of the story.
- Hugging Face BlogHugging Face ❤️ TorchTPU: TPUs, Natively
Hugging Face has released TorchTPU, making TPU a native backend device for PyTorch. Transformers, Diffusers, and Kernels can now run directly on TPU. Users only need to select a model from the Hugging Face Hub and specify device_map="tpu".