Warning: NoobAI users, lower strength to about 0.2—full support is still experimental. Seriously Consider using my settings and using Wai NSFW for this. Im working on it, Theres a bunch of other issues you can see from like my oc Lora clashes... im working on it <a rel="ugc" href="https://civarchive.com/models/915440?modelVersionId=1221116">Works well combined with phoenicians low light chiaroscuro </a> This IL model was trained on WAI-NSFW-Illustrious-SDXL and NoobAI checkpoints. Not all IL checkpoints will work as shown in my previews. Let me know in the comments if a checkpoint doesn’t work and Ill look into it. <ul><li>Optimized for creating detailed images in low-light and dark conditions... night photos</li><li>For best results... use my special settings... these were a secret for the longest time, these technically work for Sd1.5 aswell and was originally for Sd1.5</li></ul>Positive: , [:(masterpiece,best quality,very aesthetic,absurdres,newest,:1.4 ):9] Negative: transparent background, [:lowres,(bad quality,worst quality:1.2),bad anatomy,sketch,jpeg artifacts,ugly, poorly drawn,(censor),blurry,watermark,simple background:5] [(masterpiece,best quality, very aesthetic,absurdres,newest:1.4 ):2]There is still plenty to do with this one, ill be addressing them over time as I create and build up my datasetsThis upload is part of my cadence improvement strategy. I’m releasing functional but experimental LoRAs. Expect steady updates as I refine them to meet final quality standards.

[Modifier] Dark night images

Midis_Dark_V0.2[IL].safetensors

v0.2

Warning: NoobAI users, lower strength to about 0.2—full support is still experimental. Seriously Consider using my settings and using Wai NSFW for this. Im working on it, Theres a bunch of other issues you can see from like my oc Lora clashes... im working on itThis IL model was trained on WAI-NSFW-Illustrious-SDXL and NoobAI checkpoints. Not all IL checkpoints will work as shown in my previews. Let me know in the comments if a checkpoint doesn’t work and Ill look into it.<ul><li>Optimized for creating detailed images in low-light and dark conditions... night photos</li><li>For best results... use my special settings... these were a secret for the longest time, these technically work for Sd1.5 aswell and was originally for Sd1.5</li></ul>Positive: , [:(masterpiece,best quality,amazing quality,very aesthetic,absurdres,newest,:1.4 ):9] Negative: transparent background, [:lowres,(bad quality,worst quality:1.2),bad anatomy,sketch,jpeg artifacts,ugly, poorly drawn,(censor),blurry,watermark,simple background:5] [(masterpiece,best quality,amazing quality,very aesthetic,absurdres,newest:1.4 ):2]

Warning: NoobAI users, lower strength to about 0.2—full support is still experimental. Seriously Consider using my settings and using Wai NSFW for this. Im working on it, Theres a bunch of other issues you can see from like my oc Lora clashes... im working on it <a rel="ugc" href="https://civitai.com/models/915440?modelVersionId=1221116">Works well combined with phoenicians low light chiaroscuro </a> This IL model was trained on WAI-NSFW-Illustrious-SDXL and NoobAI checkpoints. Not all IL checkpoints will work as shown in my previews. Let me know in the comments if a checkpoint doesn’t work and Ill look into it. <ul><li>Optimized for creating detailed images in low-light and dark conditions... night photos</li><li>For best results... use my special settings... these were a secret for the longest time, these technically work for Sd1.5 aswell and was originally for Sd1.5</li></ul>Positive: , [:(masterpiece,best quality,very aesthetic,absurdres,newest,:1.4 ):9] Negative: transparent background, [:lowres,(bad quality,worst quality:1.2),bad anatomy,sketch,jpeg artifacts,ugly, poorly drawn,(censor),blurry,watermark,simple background:5] [(masterpiece,best quality, very aesthetic,absurdres,newest:1.4 ):2]There is still plenty to do with this one, ill be addressing them over time as I create and build up my datasetsThis upload is part of my cadence improvement strategy. I’m releasing functional but experimental LoRAs. Expect steady updates as I refine them to meet final quality standards.

Files with SHA256: 5d109e9cd2cfc74ea4ab818db547a6390cbb5ec1ec8faef847943e824c9cbc19