Try size: 1024 x 768 or other
Weights: 1.0
Note: 我重新訓練了這個Lora。它有非常多肢體、車廂結構跟內部細節上的錯誤，如果抽到奇怪的圖就當搞笑圖好了。而且如果與真實的捷運車廂照片太相似的話，可能引來不必要的麻煩。訓練台北捷運有一定程度的爭議性，畢竟它是一個高度管理的公共空間，車廂內裝的辨識度極高，也是成人影片跟主題旅館會想爭相模仿的場景。所以，希望你們盡量不要對外散布NSFW的圖片。至於生成無傷大雅的搞笑風或攝影風圖片，我覺得就隨個人自由了，因為這已是一個充斥著AI作品的世界了。素材的臉部大多數是由AI生成的，但是因為擬合度偏高，要搭配其他lora也很困難。基於上述考量，我也不打算訓練PONY或其他版本，其他人若要訓練也不干我的事。最後，如果很明顯地造成捷運公司的困擾，我得知狀況後也會配合下架的。(I've retrained this Lora. It has many errors in limbs, carriage structure, and interior details, so if weird images come up, just take them as funny pictures. Also, if the generated images look too similar to real MRT train photos, it might cause unnecessary trouble. Training on Taipei MRT is somewhat controversial since it is a highly managed public space with a very recognizable interior, and it's a scene that ***** films and themed motels would want to imitate. Therefore, I hope you all refrain from spreading NSFW images. As for generating innocuous funny or photographic pictures, I think it’s up to the individual, because this is a world full of AI works. Most of the faces in the material were generated by AI, but due to some overfitting, it was difficult to match them with other lora. Based on the above considerations, I also don't intend to train PONY or other versions, and it doesn't matter to me if others want to train them. Lastly, if it clearly causes trouble for the MRT company, I will cooperate to take it down once I learn of the situation.)
