<h3 id="heading-3">使用方法/Usage</h3>这是一个pony的画风调整模型，基于pony训练赛璐珞画风并和pony本身作差得到，目的将pony的偏欧美画风导向日漫画风。 This is a model for adjusting the art style of ponies, based on training with the art style of celluloid characters and the original pony art, with the aim of shifting the pony's predominantly Western art style towards a more Japanese anime style.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/29772dbc-b300-4297-87ed-881968b91209/width=525/29772dbc-b300-4297-87ed-881968b91209.jpeg" />由于pony美漫风强调体积感，而本模型是平面与体积感的差值，所以同样你可以添加负权重来增加体积感，减少二次元绘画的平面感。推荐权重1.2到-1.2之间，当然你也可以结合需求尝试更高或者更低的值。 Due to the emphasis on volume sensation in Western comic style ponies, and this model being the difference between flatness and volume sensation, you can add negative weights to enhance volume sensation and reduce the two-dimensional feel of the drawing.I would recommend weights ranging from 1.2 to -1.2. Of course, you can also experiment with higher or lower values based on your specific needs.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/814d02c8-5a43-4eaa-aaa8-09815da94fa5/width=525/814d02c8-5a43-4eaa-aaa8-09815da94fa5.jpeg" />基于这个原理，你同样可以用它增强需要强调体积感的模型，或是减少体积感回归原始形态。示例模型by wangka：<a target="_blank" rel="ugc" href="https://civarchive.com/models/400329/pvc-style-modelmovable-figure-model-pony">[PVC Style Model]Movable figure model Pony - Pony1.5（Light weight）FP8 | Stable Diffusion Checkpoint | Civitai</a> Based on this principle, you can use it to enhance models that require an emphasis on volume sensation or reduce volume sensation to return to the original form. Example model by wangka: <a target="_blank" rel="ugc" href="https://civarchive.com/models/400329/pvc-style-modelmovable-figure-model-pony">[PVC Style Model]Movable figure model Pony - Pony1.5（Light weight）FP8 | Stable Diffusion Checkpoint | Civitai</a><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8587c8fc-2fef-4a66-aa79-4ade42b62583/width=525/8587c8fc-2fef-4a66-aa79-4ade42b62583.jpeg" /><h3 id="heading-2664">局限性/Limitations</h3><h3 id="heading-2665">该模型是从体积感画风到平面画风的向量，训练欠拟合，微调量小，不会生成严格意义的画风，需搭配微调后的pony模型并结合实际情况使用。</h3> The model represents a vector from a volume sensation art style to a flat art style. It suffers from underfitting during training, with minimal fine-tuning. It won't strictly generate a specific art style. It needs to be used in conjunction with a fine-tuned pony model and adjusted according to the actual situation.1.可以中和诸如pvc模型搭配lora时二次元画风引入的污染，但对底模所已泛化到的角色上生成的效果不会有所改观。 It can mitigate contamination introduced by the two-dimensional art style when combined with PVC models like Lora. However, it won't significantly improve the results generated on characters that have already been generalized on the base model.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6d77bef2-7867-41be-8154-9d9c0bf8d00a/width=525/6d77bef2-7867-41be-8154-9d9c0bf8d00a.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/83db9417-3ef6-4b4d-b844-97cd16a682a7/width=525/83db9417-3ef6-4b4d-b844-97cd16a682a7.jpeg" />2.当lora对赛璐珞画风较好拟合时，其lora本身就能把pony的画风掰直，掰不直的再尝试用这个模型吧（。 When Lora fits well with the celluloid style, it can naturally align the pony's art style. If it doesn't align well, then you can try using this model to adjust it further.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/4901c3f1-1a90-4e67-b34d-97055d390244/width=525/4901c3f1-1a90-4e67-b34d-97055d390244.jpeg" />补充示例：<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/7a1ebd1d-878c-4820-a454-93ff057425ae/width=525/7a1ebd1d-878c-4820-a454-93ff057425ae.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/67171c2a-6402-4868-b92c-44f496f25f23/width=525/67171c2a-6402-4868-b92c-44f496f25f23.jpeg" />测试模型来自 <a target="_blank" rel="ugc" href="https://civarchive.com/user/Chenkin">Chenkin Creator Profile | Civitai</a> The testing model is from the <a target="_blank" rel="ugc" href="https://civarchive.com/user/Chenkin">Chenkin Creator Profile | Civitai</a>

Files with SHA256: cc6ffd5c803c3a23cee437d3bdc8f9196b79b1d6d7ab3933164c2224a099ba51