<h3 id="heading-330">Moonfall (sdXL_v10VAE)</h3><ul><li>Dataset: 1024*1024</li><li>Basic checkpoint: sdXL_v10VAEFix.safetensors</li><li>The dataset are high-quality real photos about "Moonfall".</li><li>Trigger words: space, airplane, astronauts</li><li>Suggested to generate 1024*1024 images &amp; Realistic style</li></ul>⋆｡ﾟ☁︎｡⋆｡ ﾟ☾ ﾟ｡⋆⋆｡ﾟ☁︎｡⋆｡ ﾟ☾ ﾟ｡⋆⋆｡ﾟ☁︎｡⋆｡ ﾟ☾ ﾟ｡⋆⋆｡ﾟ☁︎｡⋆｡ ﾟ☾ ﾟ｡⋆⋆｡ﾟ☁︎｡⋆｡ ﾟ☾ ﾟ｡⋆( •̯́ ^ •̯̀) ”The lora train in SDXL 1.0 cost me more than 4 hours, I feel XL have great power to generate Realistic photos than SD 1.5. This lora is trained a few epochs (20) and dataset (17P), 6800 STEPS, it is still large room to be trained. now, it is ok with weight 1.”q(╥﹏╥)p SDXL 1.0 在 txt2img 出写实摄像确实不错，但是个人觉得对设备要求比之前高了很多，这个lora我花了大4个小时练，只用了20个epochs 17张图，共6800步炼制，我其实觉得练到13k步都是可以的，目前这个lora出图在权重是1的时候还凑合，因为预感到这个下载量可能不是很多(毕竟不是nfsw,而且有设备门槛)。Hope any feedback from you, you can link me on🛸 <a target="_blank" rel="ugc" href="https://discord.gg/DkZNjeMZuV">discord</a>, <a target="_blank" rel="ugc" href="https://huggingface.co/can34/Modill">huggingface</a>, <a target="_blank" rel="ugc" href="https://github.com/cccancc">github</a>.🌸All train data collection are human-manual, and training cloud environment costs me much. hope your donation to support my new model. 🤑<a target="_blank" rel="ugc" href="https://ko-fi.com/cecily34">https://ko-fi.com/cecily34</a>If you have any questions related to the model or its samples, how to get started or just want to share your creations, feel free to join my discord:  <a target="_blank" rel="ugc" href="https://discord.gg/GgZFF3qPEB">https://discord.gg/GgZFF3qPEB</a>That is my pleasure.💕 ( •̥́ ᦷ •̀ ) Loves U.

Files with SHA256: eef2928c3faf449edbc3db670b9c385047f443652efa67a5b6bb54e7bd069098