A contrastive language-audio pretraining model using HTSAT audio encoder without fusion for zero-shot audio classification.