Update README.md

README.md

@@ -6,9 +6,14 @@ base_model:
 - Undi95/Llama-3-LewdPlay-8B
 library_name: transformers
 tags:
-- mergekit
+- not-for-all-audiences
+- nsfw
 - merge
 ---
+* <span style="color:orange">I'm just tinkering. All credit to the original creator: [Undi](https://huggingface.co/Undi95).</span>
+* <span style="color:orange">"rpcal" designates that this model was quantized using an [RP-specific data set](https://huggingface.co/datasets/royallab/PIPPA-cleaned) instead of the generalized wiki or llama data set. This is likely the last model I will create with this method as Llama-3-8B seems to get markedly dumber by doing it this way. In previous models, it was difficult to tell, but the margin of error increase from quantizing Llama-3-8B makes it obvious which method is better. I deleted the lower quants of rpcal because they are pretty dumb by comparison. This one seems to work fine, and is the only one I would recommend if you want to compare with the other, yourself. </span>
+* <span style="color:orange">This model: EXL2 @ 8.0 bpw using RP data for calibration.</span>
+---
 
 # LewdPlay-8B