Quanto: a PyTorch quantization backend for Optimum

In three linesHugging Face releases Quanto, a PyTorch quantization backend integrated into Optimum. The tool reduces model size and accelerates inference through quantization, compatible with popular transformer models.

Summary generated by Claude — human-verified