This is a major release simplifying the backends installation and improving the UX.
Major release for Speedster, now supporting TensorFlow backend for Hugging Face transformers.
The increasing compute and bandwidth demands for deploying recent Transformer models are posing a challenge in latency-sensitive applications. Therefore, there has been a focus on making Transformer models more efficient.
PyTorch 2.0 was launched in early December 2022 at NeurIPS 2022, and the main new features are performance improvements. Let's discover how PyTorch 2.0 performs against other inference accelerators.