vLLM adds engine-based parser for Plamo3 models

The vLLM repository has added an engine-based parser to support the Plamo3 model architecture. This enables native integration and efficient serving of Plamo3 within the vLLM inference runtime.