🧨 Diffusers Experimental

We are adding experimental code to support novel applications and usages of the Diffusers library. Currently, the following experiments are supported:

• Reinforcement learning via an implementation of the Diffuser model.