Add smaller model info to the readme

README.md

@@ -31,7 +31,16 @@ extra_gated_heading: Please read the LICENSE to access this model
 library_name: diffusers
 ---
 
-# Riffusion
+# Riffusion (Smaller model)
+
+This model is smaller in filesize ([2.13GB](/GitMylo/riffusion-model-v1-small) VS [14.6GB](/riffusion/riffusion-model-v1)) due to being lower precision and having ema weights and some things stuff stripped.
+* Generated spectrograms will be different from the ones in the 14.6GB model
+* There is no noticable quality difference between the original and the small model
+* The small model is easier to load on low cpu RAM, for example: If you have only 16GB of RAM, loading the large model could have some issues, like in my case, my pc froze for a few seconds.
+* The small model loads faster than the large model
+* The large model is probably better for training, but i have had great success with training LoRA on the small model.
+
+# Riffusion (Original readme)
 
 Riffusion is an app for real-time music generation with stable diffusion.