I currently have a microservice that uses Audiocraft’s MusicGen (Meta) to generate music based on text input. When I run the service on my local machine without a GPU, it takes about 9 minutes to generate the music. I’ve researched options like using AWS EC2 instances with GPUs, Docker with NVIDIA GPU support, etc., but I’m unsure which is the best choice for a beginner who doesn’t want to spend too much money on this side project. Any recommendations on how to speed up the process affordably and effectively?