Version naming info: v1.0 BM-1.41 LyCORIS r128 Version of this model BoleroMix-version Model Type Network Dimension/Rank of the modelThis is a LyCORIS extract of <a target="_blank" rel="ugc" href="https://civarchive.com/models/448716">BoleroMix</a> by <a target="_blank" rel="ugc" href="https://civarchive.com/user/bolero537">bolero537</a>By using <a target="_blank" rel="ugc" href="https://github.com/hako-mikan/sd-webui-lora-block-weight">LoRA Block Weight</a> (<a target="_blank" rel="ugc" href="https://github.com/ltdrdata/ComfyUI-Inspire-Pack">Inspire-Pack</a> has LBW nodes for Comfy) and this LyCORIS I can use BoleroMix's style on other checkpoints (gallery has comparison grids) Upd.: I generated some images with this and the quality is pretty good, i might use them to make a style LoRA<h3 id="for-lycoris-lbw-use-00000000000111111111-to-apply-style-to-slightly-enhance-the-effect-use-weight-1-or-111111111111111111111-(comparison-image)-dod15xkir">For LyCORIS LBW use <code>0,0,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1</code> to apply style; to slightly enhance the effect use weight = 1 or <code>1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1</code> (<a target="_blank" rel="ugc" href="https://civarchive.com/images/34651812">comparison image</a>)</h3>I created this because I really like the aesthetic of BoleroMix and also like the <a target="_blank" rel="ugc" href="https://civarchive.com/models/282341">4th Tail model</a> and I wanted BoleroMix's style on 4th Tail. The problem with it, or rather me, is that I don't know any artists that make up a similar style and it's just much more convenient if the model has a nice stable base style I know and like overall instead of switching around artist styles I picked out (might upload my wildcard file at some point). And a plus point is that it can be used on other models too, and probably on the on-site generator Info: <code>0,0,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1</code> will only apply the OUT layers (identical to the OUT12 block merge preset) of the LyCORIS model , in theory it applies all aesthetic/style from the model to a checkpoint but weight = 1 or setting all layers to 1 will get it closer to the original model of the extract while possibly losing composition/concept knowledge of the checkpoint its applied to (im not sure about this so correct me if im wrong).<a target="_blank" rel="ugc" href="https://github.com/hako-mikan/sd-webui-lora-block-weight">SD WebUI LBW </a>preset: FXL_OUT12:0,0,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1 Usage: <code>&lt;lora:BoleroMix_LycoExtract_r128:1:1:lbw=0,0,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1&gt;</code> <code>&lt;lora:BoleroMix_LycoExtract_r128:1:1:lbw=FXL_OUT12&gt;</code>Note: The comparison grids (except the simple 1x2) have the Comfy workflow embedded into them.Info on the extract: This extract was made using Kohya_ss GUI in the LoRA &gt; Tools &gt; Extract LyCORIS LoCon tab and used PonyXL v6 for the base model selectionI may try around with lower rank extracts and see how it affects the style, so far there wasnt much difference in style in comparison to higher rank, so far with the existing extracts (see comparison grid) i dont see much difference in style and that is why i uploaded the lowest rank lycoris. Also as already mentioned: I generated some images with this and the quality is pretty good, i might use them to make a style LoRA (Information even if I haven't uploaded a LoRA version: XL LoRA has less layers so the equivalent would be <code>0,0,0,0,0,0,1,1,1,1,1,1</code>, ALSO: if you see LBW stuff online and they have 17 LoRA/26 LyCORIS layers, they are for sd1.5. sd1.5 models have more layers than sdxl)If the author (bolero) wants me to remove this model because it doesn't count as merge I will do so. This model or merges are not allowed to be uploaded anywhere outside of CivitAI and merges on CivitAI must use the same license as this model/<a target="_blank" rel="ugc" href="https://civarchive.com/models/448716">BoleroMix</a>If you know any artists that have a similar style to BoleroMix please leave a comment/DM me :&gt;

BoleroMix LyCORIS Extract

Network/Conv dimension (rank) of 128This extract was made using Kohya_ss GUI in the LoRA &gt; Tools &gt; Extract LyCORIS LoCon tab and used PonyXL v6 for the base model selection. Mode: fixed; net dim 128; conv dim 128 no sparse bias

Files with SHA256: e9761c1b5cf31bba2a1dcec0a40403bea85f842b3934d007bf3fd8fc7210a601