Try size: 768 x 1024 or 960 x 1280 or 1024 x 1024 (ILXL and OOO)Weights: 0.8~1.0Clip skip: 2Pictures from kantanmix_v16LCMTurbo and OOO_v80Sampling steps: 20~40 (It is suggestion 40 in duel / 雙人對決建議40)Note: Note that the basic model I used to generate the image is LCM/Turbo, so the total number of steps and CFG are less. But in the same way, the two characters will inevitably make mistakes, and increasing the number of steps will reduce the error rate. It is easy to happen Tsuuhaiken. If you don’t like it, you can delete the trigger words, or strengthen the other posture description, or reduce the LoRA weight. / 注意我的生成圖片的基礎模型是LCM/Turbo，所以總步數跟CFG是比較少的。但同理，兩位角色難免會出錯，提高步數會使出錯率降低一些。很容易出現通背拳，假如你不喜歡可以刪除觸發詞，或加強其他姿勢描述，或降低LoRA權重。(ILXL)Note: It has a lot of weird bugs, so it's not for perfectionists. But if you have "Tsuuhaiken Addiction Syndrome", you can give it a try. <s>I'm not sure if there will be an ILXL version yet, as it's hard to train...</s> / 它有很多奇怪的錯誤，所以不適合完美主義者。但如果你得了"通背拳上癮症候群"，那你可以試玩看看。<s>我目前還不確定會不會有ILXL版本，因為這個很難訓練... </s>(OOO)Reference Video:<div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/qGMixvlopmw" start="0"></iframe></div>

Chinmi & Tsuuhaiken (Tekken Chinmi) / チンミ & 通背拳 (鉄拳チンミ)

Chinmi&Tsuuhaiken(TekkenChinmi)_anyilxl011.safetensors

ILXL

Try size: 768 x 1024 or 960 x 1280 or 1024 x 1024 (ILXL and OOO)Weights: 0.8~1.0Clip skip: 2Pictures from kantanmix_v16LCMTurbo and OOO_v80Sampling steps: 20~40 (It is suggestion 40 in duel / 雙人對決建議40)Note: Note that the basic model I used to generate the image is LCM/Turbo, so the total number of steps and CFG are less. But in the same way, the two characters will inevitably make mistakes, and increasing the number of steps will reduce the error rate. It is easy to happen Tsuuhaiken. If you don’t like it, you can delete the trigger words, or strengthen the other posture description, or reduce the LoRA weight. / 注意我的生成圖片的基礎模型是LCM/Turbo，所以總步數跟CFG是比較少的。但同理，兩位角色難免會出錯，提高步數會使出錯率降低一些。很容易出現通背拳，假如你不喜歡可以刪除觸發詞，或加強其他姿勢描述，或降低LoRA權重。(ILXL)Note: It has a lot of weird bugs, so it's not for perfectionists. But if you have "Tsuuhaiken Addiction Syndrome", you can give it a try. <s>I'm not sure if there will be an ILXL version yet, as it's hard to train...</s> / 它有很多奇怪的錯誤，所以不適合完美主義者。但如果你得了"通背拳上癮症候群"，那你可以試玩看看。<s>我目前還不確定會不會有ILXL版本，因為這個很難訓練... </s>(OOO)Reference Video:<div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/qGMixvlopmw" start="0"></iframe></div>

Files with SHA256: cc084388bf30551b31cdeeb616c0135609c13f7b55f8c7d33cdcb6599778a701