This cluster only has 2vCPU, which is not sufficient to transcribe an audio snippet of that size fast enough. So you want to either reduce the audio snippet size, or add more CPU cores