vLLM Adds FP8 Quantization Support for ModernBERT

vLLM now supports FP8 quantization for ModernBERT models. This addition enables more efficient inference and reduced memory footprint for the ModernBERT architecture within the runtime.