<h3 id="guidance:">Guidance:</h3><ul><li>This lora can be used with t2v or i2v.</li><li>The v2 version can possess the target in physical or ghost form.</li><li>In the "About this version" section you can find a list of the most frequent nouns and verbs in the training set.</li><li>If the possession effect does not work, consider increasing the high-noise LORA; if there is contamination to other detail-enhancing LORAs, try lowering the low-noise LORA.</li></ul><h3 id="preview-image-prompt:">Preview Image Prompt:</h3>Possession. A man enters a woman's body and possesses her. The woman experienced pain and severe tremors after being possessed, but then returned to normal.<h3 id="version:">Version:</h3><ul><li>v2: Expand and clean the dataset, then adjust the training parameters and retrain with the help of Gemini Pro.</li><li>v1: Attempted to train my first LORA, using Smiles_wx porting trainer, the Windows version of diffusion-pipe, from <a target="_blank" rel="ugc" href="https://www.bilibili.com/video/BV1dJtFz2EGD/">https://www.bilibili.com/video/BV1dJtFz2EGD/</a></li></ul><h3>说明：</h3><ul><li>这个lora可使用在t2v或是i2v。</li><li>v2 版本可以以实体或幽灵形态附身目标。</li><li>在"About this version"部分，你可以看到训练集中频率最高的名词与动词清单。</li><li>如果没有附身效果，可以考虑增加高噪LORA；如果对其他细节增强型 LORA 造成污染，请尝试降低低噪LORA。</li></ul><h3>预览图提示词：</h3>附身。一名男子进入一名女子体内并附身。该女子被附身后感到疼痛和剧烈颤抖，但随后恢复正常。<h3>版本：</h3><ul><li>v2：扩展并清理资料集，然后借助 Gemini Pro调整训练参数并重新训练。</li><li>v1：尝试使用 Smiles_wx 整合的训练器（Windows 版的 diffusion-pipe）训练我的第一个 LORA，该训练器来自 <a target="_blank" rel="ugc" href="https://www.bilibili.com/video/BV1dJtFz2EGD/">https://www.bilibili.com/video/BV1dJtFz2EGD/</a></li></ul>

Possession 附身

The 15 most frequent nouns in the dataset:<ul><li>woman - 63 times</li><li>body - 44 times</li><li>possession - 39 times</li><li>spirit - 39 times</li><li>light - 15 times</li><li>shock - 14 times</li><li>takeover - 14 times</li><li>control - 14 times</li><li>chest - 13 times</li><li>ghost - 12 times</li><li>energy - 9 times</li><li>man - 9 times</li><li>expression - 8 times</li><li>force - 7 times</li><li>office - 6 times</li></ul>The 15 most frequent verbs in the dataset:<ul><li>confirms - 11 times</li><li>dives - 11 times</li><li>merges - 7 times</li><li>signaling - 7 times</li><li>leaving - 7 times</li><li>jolted - 7 times</li><li>approaches - 6 times</li><li>causing - 6 times</li><li>becomes - 5 times</li><li>taken - 5 times</li><li>replaced - 5 times</li><li>runs - 4 times</li><li>seizes - 4 times</li><li>Melts - 4 times</li><li>Materializes - 4 times</li></ul>

Files with SHA256: cc4cb5c1e1acb70a443cecb34d028336fec3d8fdad37478f21a08f6c5620436a