<h2 id="yet-another-pony-merge!-2qntjikik">Yet Another Pony Merge!</h2>Merge of <a target="_blank" rel="ugc" href="https://civarchive.com/models/341433?modelVersionId=511677">Everclear PNY</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/372465/pony-realism?modelVersionId=534642">Pony Realism</a>. Originally had no intent to upload, but it seems to work better with my <a target="_blank" rel="ugc" href="https://civarchive.com/models/536317/fursuit-head-lora-for-ponyxl">Fursuit LoRA</a>, so I thought I would upload for the benefit of those who generate using Civitai's services.<h2 id="version-2.0-fq3bqxnu7">VERSION 2.0</h2><h3 id="are-you-ready-for-a-massive-improvement-zpg1hoic1">Are You Ready for a Massive Improvement?</h3>Despite my personal enjoyment of the original model, people did have some valid criticisms, namely the loss of important details over Pony Realism.To try to remedy this, this new version was weighted block-by-block over the course of several hours with a number of intentions in mind:<ol><li>Important details should be retained without having the "deepfried" look that Realism often has,</li><li>Prompt adherence should be of maximum priority,</li><li>Realistic, believable aesthetics should be as high as is possible,</li><li>Colors should be vivid without being oversaturated, and</li><li>Lora compatibility should be as high as possible</li></ol>I consider myself rather humble, but must say that I am extremely pleased with the results!2 versions of this updated model are included: v2a is the base re-weight, whilst v2b is v2a merged 1:1 with v1.v2a produces better results in 99% of cases, but in the rare case of things like prompt incompatibility of differences, v2b will be an improvement over v1.Model weights are as follows:<ul><li>Model 1: Everclear PNY v3</li><li>Model 2: Pony Realism v21</li></ul><ul><li>Time Embed: 0.15</li><li>Label Embed: 0.5</li><li>Input 0: 0.1</li><li>Input 1: 0.2</li><li>Input 2: 0.5</li><li>Input 3: 0.05</li><li>Input 4: 0.15</li><li>Input 5: 0.9</li><li>Input 6: 0.2</li><li>Input 7: 0.2</li><li>Input 8: 0.85</li><li>Middle 0: 0.5</li><li>Middle 1: 0.75</li><li>Middle 2: 0.6</li><li>Output 0: 0.85</li><li>Output 1: 0.55</li><li>Output 2: 0.9</li><li>Output 3: 0.85</li><li>Output 4: 0.75</li><li>Output 5: 0.95</li><li>Output 6: 0.6</li><li>Output 7: 1</li><li>Output 8: 0.65</li><li>Out: 0.95</li></ul><ul><li>CLIP: 0.18</li></ul>Happy generating!<h2 id="version-1.0-6dwkrw0k2">VERSION 1.0</h2>Reasoning behind the merge was that I prefer the prompt adherence and image quality of Realism, but prefer the realistic look of Everclear, and so decided to merge them such that the model weights were primarily Everclear, and the CLIP was primarily Pony Realism:Model Weights:Everclear : Realism - 0.85 : 0.15CLIP WeightsEverclear : Realism - 0.15 : 0.85In my personal opinion this seems to result in the best of both worlds, but your results and tastes may vary.

Files with SHA256: 69cd4cf93dc88c4c247152de3ad722822c9d28512116b6a714f9307dae7a644a