<h3>A LORA trained on official artworks of all the 1008 Pokémon, along with their types and BLIP generated captions.</h3>This is my first fully trained LORA model and I have no idea what I'm doing.Any feedback or tips on how to improve this would be really helpful :)Fell free to experiment with it and share!<h3>Recommended Configs:</h3>My default prompt is:pokemon, [type1] and [type2], [characteristics], masterpiece, high quality, best quality, high-definition, ultra-detailed, black background,&lt;lora:pokemon:0.5&gt;Negative prompt:nsfw, human, 1boy, 1girl, (worst quality, low quality:1.4), ( jpeg artifacts:1.4), (depth of field, bokeh, blurry, film grain, chromatic aberration, lens flare:1.0), greyscale, monochrome, dusty sunbeams, trembling, motion lines, motion blur, emphasis lines, text, title, logo, signature<ul><li>The weights could be between 0.5-0.7 for animal like and 0.2-0.5 for humanoid like Pokémon.</li><li>Clip Skip: 2 and ENSD: 31337.</li><li>768x768 works better than 512x512.</li><li>For sampler Euler a works fine, but I prefer DPM++ SDE Karras.</li><li>Use "black background" tag to keep the Pokémon isolated, more complex backgrounds are better with negative embeddings.</li><li>You can use either one or two types.</li><li>Describe your Pokémon along with vibe characteristics (Ex: edgy, cute, strong, etc).</li><li>I tested it on both <a target="_blank" rel="ugc" href="https://civarchive.com/models/9942/abyssorangemix3-aom3">AbyssOrangeMix3</a>, and <a target="_blank" rel="ugc" href="https://civarchive.com/models/7371/rev-animated">ReV Animated</a> and get good results, realistic models such as <a target="_blank" rel="ugc" href="https://civarchive.com/models/4201/realistic-vision-v13-fantasyai">Realistic Vision</a> seems to not work at all, but further tests are welcomed.</li><li>Abyss has more official artwork style and ReV is more stylized.</li><li>You can get deformed creatures some times, so a negative embedding should help.</li></ul>Examples of negative embeddings:<img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8ed99c80-b2d1-476d-4021-a9dd2b39df00/width=525/8ed99c80-b2d1-476d-4021-a9dd2b39df00" /><img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/fed242f2-a623-48b3-ec2a-db449f1a2900/width=525/fed242f2-a623-48b3-ec2a-db449f1a2900" />Here is a image with every type applied in the same prompt: <a target="_blank" rel="ugc" href="https://imgur.com/a/pkqj5ST">https://imgur.com/a/pkqj5ST</a>The base model used for the training is the <a target="_blank" rel="ugc" href="https://civarchive.com/models/9942/abyssorangemix3-aom3">AbyssOrangeMix3</a>, therefore the <a target="_blank" rel="ugc" href="https://huggingface.co/WarriorMama777/OrangeMixs#licence">licenses </a>of this LORA are the same.

Fakemon (Pokémon LORA) Beta

First version of the LORA trained at 10 Epochs and 14,400 steps.

pokemon.safetensors

v1.0 beta

<h3>A LORA trained on official artworks of all the 1008 Pokémon, along with their types and BLIP generated captions.</h3>This is my first fully trained LORA model and I have no idea what I'm doing.Any feedback or tips on how to improve this would be really helpful :)Fell free to experiment with it and share!<h3>Recommended Configs:</h3>My default prompt is:pokemon, [type1] and [type2], [characteristics], masterpiece, high quality, best quality, high-definition, ultra-detailed, black background,&lt;lora:pokemon:0.5&gt;Negative prompt:nsfw, human, 1boy, 1girl, (worst quality, low quality:1.4), ( jpeg artifacts:1.4), (depth of field, bokeh, blurry, film grain, chromatic aberration, lens flare:1.0), greyscale, monochrome, dusty sunbeams, trembling, motion lines, motion blur, emphasis lines, text, title, logo, signature<ul><li>The weights could be between 0.5-0.7 for animal like and 0.2-0.5 for humanoid like Pokémon.</li><li>Clip Skip: 2 and ENSD: 31337.</li><li>768x768 works better than 512x512.</li><li>For sampler Euler a works fine, but I prefer DPM++ SDE Karras.</li><li>Use "black background" tag to keep the Pokémon isolated, more complex backgrounds are better with negative embeddings.</li><li>You can use either one or two types.</li><li>Describe your Pokémon along with vibe characteristics (Ex: edgy, cute, strong, etc).</li><li>I tested it on both <a target="_blank" rel="ugc" href="https://civitai.com/models/9942/abyssorangemix3-aom3">AbyssOrangeMix3</a>, and <a target="_blank" rel="ugc" href="https://civitai.com/models/7371/rev-animated">ReV Animated</a> and get good results, realistic models such as <a target="_blank" rel="ugc" href="https://civitai.com/models/4201/realistic-vision-v13-fantasyai">Realistic Vision</a> seems to not work at all, but further tests are welcomed.</li><li>Abyss has more official artwork style and ReV is more stylized.</li><li>You can get deformed creatures some times, so a negative embedding should help.</li></ul>Examples of negative embeddings:<img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8ed99c80-b2d1-476d-4021-a9dd2b39df00/width=525/8ed99c80-b2d1-476d-4021-a9dd2b39df00" /><img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/fed242f2-a623-48b3-ec2a-db449f1a2900/width=525/fed242f2-a623-48b3-ec2a-db449f1a2900" />Here is a image with every type applied in the same prompt: <a target="_blank" rel="ugc" href="https://imgur.com/a/pkqj5ST">https://imgur.com/a/pkqj5ST</a>The base model used for the training is the <a target="_blank" rel="ugc" href="https://civitai.com/models/9942/abyssorangemix3-aom3">AbyssOrangeMix3</a>, therefore the <a target="_blank" rel="ugc" href="https://huggingface.co/WarriorMama777/OrangeMixs#licence">licenses </a>of this LORA are the same.

Files with SHA256: 1b49190e079116bf0bdcab5a1b594ffd1551bf86ae26881c54b8b6fbdb4bceba