VCRScore: Image captioning metric based on V\&L Transformers, CLIP, and precision-recall