Hugging Face's transformers library now powers vLLM inference at native speed, closing the performance gap between the two popular serving stacks.
Continue to AI University →