Create README.md

This is a basic UNet diffusion model that generates spoken digits (0-9) trained on AudioMNIST for 750,000 steps. No inference file yet, but I have uploaded the model weights and the training file with all parameters I used for training.

README.md

@@ -0,0 +1,5 @@
+---
+datasets:
+- flexthink/audiomnist
+pipeline_tag: text-to-speech
+---