Update README.md

README.md

@@ -2,7 +2,15 @@
 license: llama2
 ---
 
-The refined, perfected version of MythoMaxKurisu-13b. All the errors in the training code are gone, and the model now understands *actions*. Training on the dataset no longer has the tradeoff of making the model stupider, while at the same time improving its tone of voice.
+A refined version of MythoMaxKurisu-13b. All the errors in the training code are gone, and the model now understands *actions*. Training on the dataset no longer has the tradeoff of making the model stupider, while at the same time improving its tone of voice.
 
-I don't have time to write a proper model card right now (I'll come back to this). For details, refer to the MythoMaxKurisu repo, this is basically that but better.
-For quants, look on my account, you'll find them.
+This is a model specialized in roleplaying as a single character: Makise Kurisu from Steins;Gate. However, in testing the model has shown it's capable of roleplaying as other characters too.
+
+Example output (1 epoch):
+[mythomakise_v2.png]
+
+Quant (1 epoch only right now):
+https://huggingface.co/Heralax/MythoMakise-13b-quants/tree/main
+
+Adapters:
+https://huggingface.co/Heralax/MythoMakise-13b-adapters/tree/main