Whether it's prompt or not affects the results generated？

[tanggang1997]

Due to the prompt, does it happen that I would have wanted to generate sound effects, but instead I generate music, because the quality of the prompt affects the generated result a lot, or does it correspond to a specific task where I can choose the corresponding model? It's in your readme, but it doesn't seem to be an option in the huggingface api!

[tanggang1997]

sorry，i understand what you say,but does this happen with audioldm2??