<h1 id="heading-423">MODILL PASTELL - Modern Style Illustration LoRa</h1>This LoRa the more "pastel/detailed" one. Still my biggest LoRa yet, ~500 images for a bit less than 50000 steps x2 (I trained it two times with different settings and merged the results) and then carefully merged with other unpublished LoRa of mine. As like the other, Its dataset contains mainly modern style illustration+the dataset prevoiusly used for CoolKids LoRa and the datasets for the others LoRa I mixed in. It should be more flexible (especially at adults themed images) and it should work on a wide variety of subjects.<h2 id="heading-332">!!! IMPORTANT - READ BEFORE LEAVING REVIEWS, PLEASE !!!</h2>In the example images the weights are not present since the LoRa was applied using the additional network plugin in A1111 instead of the native LoRa extension. Below are the suggested weights, tune them until your images comes out without distortions!Use it at ~0.3 weight if you use it on a illustration like checkpoints (Dreamshaper is a good example) for a digital illustration lookUse it at ~0.45 up to 0.6 weight if you use it on a realistic like checkpoints (Realistic Vision is a good example) for a pastell illustration lookI also found that using the word "whimsical" yeld pretty good results, consider incorporating that into your promptsTraining LoRa necessitate a lot of coffee. If you think that what I do is useful and want to help me keep doing this you can buy me a coffe here: <a target="_blank" rel="ugc" href="https://ko-fi.com/clumsy_trainer">https://ko-fi.com/clumsy_trainer</a>

MODILL PASTELL - Modern Style Illustration LoRa

MODILL_PASTEL_RC_V0.01.safetensors

v1.0

MODILL PASTELL - Modern Style Illustration LoRaThis LoRa the more "pastel/detailed" one. Still my biggest LoRa yet, ~500 images for a bit less than 50000 steps x2 (I trained it two times with different settings and merged the results) and then carefully merged with other unpublished LoRa of mine. As like the other, Its dataset contains mainly modern style illustration+the dataset prevoiusly used for CoolKids LoRa and the datasets for the others LoRa I mixed in. It should be more flexible (especially at adults themed images) and it should work on a wide variety of subjects.!!! IMPORTANT !!!In the example images the weights are not present since the LoRa was applied using the additional network plugin in A1111 instead of the native LoRa extension. Below are the suggested weights, tune them until your images comes out without distortions!Use it at ~0.3 weight if you use it on a illustration like checkpoints (Dreamshaper is a good example) for a digital illustration lookUse it at ~0.45 up to 0.6 weight if you use it on a realistic like checkpoints (Realistic Vision is a good example) for a pastell illustration look I also found out that using the word "whimsical" yeld pretty good results, consider incorporating that into your promptsTraining LoRa necessitate a lot of coffee. If you think that what I do is useful and want to help me keep doing this you can buy me a coffe here: https://ko-fi.com/clumsy_trainer

Files with SHA256: 34fc8a925ffa408fc3c2354b7b37df2212a59fb6fd14e3903c8de3214ba5c3fd