こちらがT2

で、おなじプロンプトを用いてLoRAも使って作ったのがこちらの作品(∩´∀｀)∩　プロンプトに書いてある情報通りなのですが、面白い事に自分が意図したことに近かったのはSDモデルの方です。　T2は情報を読み込みすぎて仕上げられた感じですｗ　ただやはり安定感はＴ２の方がすごいあります（当たり前ですがｗ）　もっと自分にＴ２を使える技術があれば、多分ＳＤの様な感じにも出来ると思います(ノД`)・゜・。　 So, here is the piece I created using the same prompt along with LoRA. It follows the prompt exactly, but interestingly enough, the standard SD model actually produced something closer to what I had originally envisioned. The T2 version feels like it tried to incorporate a bit too much of the prompt information into the final result. That said, T2 definitely offers superior stability (which is to be expected, of course). If I had the technical skills to master T2, I’m sure I could achieve a look similar to the SD version.