Assessing the alignment between infants' visual and linguistic experience using multimodal language models