From the course: Hands-On Generative AI with Diffusion Models: Building Real-World Applications
Join today to access over 25,300 courses taught by industry experts.
Beyond images with diffusion models
From the course: Hands-On Generative AI with Diffusion Models: Building Real-World Applications
Beyond images with diffusion models
- [Instructor] It's time now to delve into the fascinating world of diffusion models beyond images. These models are known for their capabilities that encompass fields like video generation, protein synthesis, and even reinforcement learning. So let's dive right in. The first use case that we have is audio generation. AudioLDM is a good example of a significant leap in text to audio generation systems. This model offers high quality audio synthesis just based on text descriptions. The system involves a latent space learning continuous audio representations from clap latent. The pre-trained clap models facilitate training latent diffusion models with audio embeddings easy. AudioLDM excels in both generation quality and computational efficiency. Unlike previous audio generation systems, it enables text guided audio manipulations like style transfer, achieving state of the art performance on audio. Next up, we have diffusion models in reinforcement learning. In reinforcement learning…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.