This DORA is designed to recreate the style of the artist A6 (aka eeroku). Read the usage notes for details on each version. A HiRes fix pass is recommended for all versions for best results.NSFW NOTE: Due to the checkpoints it is trained on, this DORA is very NSFW capable. If you do not want NSFW make sure you actively prompt away from it. Also note that NoobAI tends to age-down characters, so make sure to prompt accordingly to avoid violating the CivitAI TOS when generating NSFW.All sample images were created directly out of ReForge UI with a Hires Fix pass (I suggest Latent Exact Nearest with ~0.55-0.68 denoise). No inpainting or ADetailer was used. You may want to try "sketch" or "pencil outline" in the negative to get smoother outlines with a latent upscale.Please support the original artist wherever possible so they can continue to produce their art. Without them, LORAs such as these would not be possible.This model will also backed up on Huggingface and is available on TensorArt: <a target="_blank" rel="ugc" href="https://tensor.art/models/788960212368443662/">https://tensor.art/models/788960212368443662/</a>Please do not violate the content policy -&gt; <a target="_blank" rel="ugc" href="https://civarchive.com/safety#content-policies">https://civarchive.com/safety#content-policies</a>Please do not post NSFW images of minors to this model page<h3 id="usage-notes:">USAGE NOTES:</h3>For both versions, it is recommended to use it at full strength. If you are using a different checkpoint than it was trained on, you may need to increase the strength to retain the style. Both versions were trained with an additional tag of "pencil outline" that you may want to put in the negative if you are having issues with HiRes fix.PONY:This LORA was trained on AutismMix. Most images were tagged with "source_anime" and scores 6-9.Recommended Positive prompt:<pre><code>score_9, score_8_up, score_7_up, score_6_up, source_anime</code></pre>Recommended Negative prompt:<pre><code>source_pony, (hands:1.1)</code></pre>NoobAI:The NoobAI version was trained on NoobAI Epsilon v1.1. This should be compatible with NoobAI V-Pred, as well, as you can see in the sample gallery, just make sure you adjust your sampler/CFG accordingly. I won't be making a separate V-pred version because I feel like the added contrast of V-Pred isn't consistent with the artist's actual style. For NoobAI, you must use the trigger word "eeroku-style"Recommended Positive prompt:<pre><code>masterpiece, best quality, eeroku-style</code></pre>Recommended Negative prompt:<pre><code>worst quality, bad quality, screenshot, sketch, jpeg artifacts</code></pre><h3 id="tagging-notes:">TAGGING NOTES:</h3>The dataset contained over 370 images that were ranked.Score 9-7/"masterpiece," "best quality," "good quality" images were the most recent and detailed, while score 7/"good quality" were older images or lower detail.Score 6/"average quality" images are the least detailed and also include some images sourced from manga scans that have been tagged with "source_manga"/"manga" and "monochrome, screentones" tags.

A6 (eeroku) Style for NoobAI/Pony

a6EerokuStyleFor_v1nbNoobaiEps.safetensors

v1Nb [NoobAI Eps]

v1g [Pony]

This DORA is designed to recreate the style of the artist A6 (aka eeroku). Read the usage notes for details on each version.HiRes Fix is recommended for the best results. Most sample images were generated with 1.5x Latent (Nearest) upscale with ~20 steps and 0.58-0.7 denoise. All samples were generated directly from ReForge with no use of inpainting or Adetailer. You may want to try "sketch" or "pencil outline" in the negative to get smoother outlines with a latent upscale.Please support the original artist wherever possible so they can continue to produce their art. Without them, LORAs such as these would not be possible.This model is also backed up on Huggingface and CivitAI: <a href="https://civitai.com/models/890946/a6-eeroku-style" rel="ugc noopener noreferrer nofollow" target="_blank">https://civitai.com/models/890946/a6-eeroku-style</a>USAGE NOTES:For both versions, it is recommended to use it at full strength. If you are using a different checkpoint than it was trained on, you may need to increase the strength to retain the style. Both versions were trained with an additional tag of "pencil outline" that you may want to put in the negative if you are having issues with HiRes fix.PONY:This LORA was trained on AutismMix. Most images were tagged with "source_anime" and scores 6-9.Recommended Positive prompt:score_9, score_8_up, score_7_up, score_6_up, source_animeRecommended Negative prompt:source_pony, (hands:1.1)NoobAI:The NoobAI version was trained on NoobAI Epsilon v1.1. This should be compatible with NoobAI V-Pred, as well, as you can see in the sample gallery, just make sure you adjust your sampler/CFG accordingly. For NoobAI, you must use the trigger word "eeroku-style"Recommended postive prompt:masterpiece, best quality, eeroku-styleRecommended negative prompt: (you can add "hands" and "fingers" as needed)worst quality, bad quality, screenshot, sketch, jpeg artifactsTAGGING NOTES:The dataset contained over 370 images that were ranked. Most dataset images also included a certain Japanese word for male youths as a tag, so that may contain some of the style data, as well, but it also makes the model more flexible.Score 9-7/"masterpiece," "best quality," "good quality" images were the most recent and detailed, while score 7/"good quality" were older images or lower detail.Score 6/"average quality" images are the least detailed and also include some images sourced from manga scans that have been tagged with "source_manga"/"manga" and "monochrome, screentones" tags.

Files with SHA256: 4e3b985e975a39b7aecee65dc6b432e01d138623c69d33c05b5dd00de94324cb