GhostXL is an anime model. If you like GhostXL, please give it a 5 stars review. It is very inspiring for me to improve GhostXL.(中文版本在liblib和tusi都有)Any questions or corporations? Find me at <a target="_blank" rel="ugc" href="https://discord.gg/576hznSTS7">Discord</a> or by Wechat: ghostinshell10. ThanksHello everyone, I am GhostInShell, the author of GhostMix. First and foremost, I would like to express my sincerest thanks to each and every user of GhostMix. With so many models out there, it is my honor that you guys choose and love GhostMix. Once again, thank you all. Secondly, I sincerely apologize for the long delay of GhostXL. Because I have high standards for the model and strive for perfection, a lot of time has been spent training and adjusting the model. Although GhostXL-V1.0 is not as good as I had imagined, it has a unique style and meets my standards in various aspects, therefore I decided to release it.Before using GhostXL, I recommend reading the user instructions below. Thank you.<h3 id="heading-121">User Instructions for GhostXL-V1.0</h3><ol><li>After extensive testing, we found that GhostXL-V1.0 can directly generate images with high precision in the resolution range of 1024-1536, and there's no need for upsampling or refining operations.</li><li>xx art style: The "xx art style" option allows for fine adjustments of the style. Here are some examples illustrating the effects of different art styles with the same prompts and seeds. (PS: If you want effects similar to GhostMix, the "realistic" or "realistic anime" art styles are personally recommended.)<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/adc0a143-bd8a-43c1-b0d8-b614cb5ddd31/width=525/adc0a143-bd8a-43c1-b0d8-b614cb5ddd31.jpeg" /></li><li>Personal observation: The compatibility of GhostXL seems slightly less than GhostMix, and its image quality might be a bit inferior, although it tends to produce a higher percentage of satisfactory results. (These comparisons are made against an upscaled version of GhostMix.) The primary reason for releasing GhostXL is the current market's lack of models with this particular style. On the bright side, GhostXL has an edge in producing immediate outputs, making upsampling unnecessary.</li></ol><h3 id="heading-2">GhostXL Model Details</h3><ul><li>Developed by: <a target="_blank" rel="ugc" href="https://civarchive.com/user/_GhostInShell_">GhostInShell</a></li><li>Model type: Diffusion-based text-to-image generative model</li><li>License: <a target="_blank" rel="ugc" href="https://freedevproject.org/faipl-1.0-sd/">Fair AI Public License 1.0-SD</a></li><li>Finetuned from model: <a target="_blank" rel="ugc" href="https://huggingface.co/Linaqruf/animagine-xl-2.0">Animagine XL 3.0</a></li><li>Config of Finetuning: <a target="_blank" rel="ugc" href="https://github.com/drnighthan/GhostXL/blob/main/GhostXL_finetune.toml">GhostXL_finetune.toml</a></li><li>OtherMergeModel: <a target="_blank" rel="ugc" href="https://civarchive.com/models/43977/leosams-helloworld-xl?modelVersionId=338512">LEOSAM's HelloWorld XL:https://civarchive.com/models/43977/leosams-helloworld-xl?modelVersionId=338512</a><a target="_blank" rel="ugc" href="https://civarchive.com/models/112902?modelVersionId=351306">DreamShaper XL: https://civarchive.com/models/112902?modelVersionId=351306</a></li></ul><h3 id="heading-122">Analysis of the Pros and Cons of GhostXL-V1.0</h3>As I mentioned in my previous articles, the main factors I consider when creating models are: 1) Compatibility; 2) Image quality; 3) Rate of generating satisfactory results.<ol><li>Compatibility: As a model that generates images in a 2D-2.5D style, GhostXL-V1.0 shows lesser compatibility with different styles when compared to GhostMix, especially for 3D styles. GhostXL is a finetune of Animagine XL V3, combined with features from other models (70% base from the finetuned model and 30% from the blend of others). As of the 1.0 version, GhostXL is still primarily a 2D-2.5D model, but I plan to keep developing it to see whether the style compatibility can be improved. For prompts compatibility, thanks to Animagine XL V3, GhostXL gives a good response to prompts such as 'masterpiece', 'best quality', as well as anime characters. However, it doesn't perform as well with more stylized prompts.</li><li>Image Quality: Personally, I feel that the image quality is generally satisfactory, considering that GhostXL generates images in the range of 1024-1536 directly (one-stage generation), unlike GhostMix which generates in 512-768 and then undergoes upsampling (a two-stage generation process). I am quite satisfied with the quality it achieves using the one-stage approach.</li><li>The Rate of Generating Satisfactory Results: In my experience, I believe GhostXL yields a higher rate of satisfactory results than GhostMix. However, for the same prompts, the variation in results is less broad in GhostXL relative to GhostMix. To increase diversity within the same theme, prompts may need to be adjusted.</li></ol>

GhostXL

ghostxl_v10BakedVAE.safetensors

v1.0-Baked VAE

Hello everyone, I'm the author of GhostMix _GhostInShell_. First, thank you very much to every GhostMix user. It is my honor to be able to use and love GhostMix among so many models. Second, I'm very sorry that GhostXL has been playing for so long, because I have high requirements for the model, and I have to do my best to do it, so I spent a lot of time training and adjusting the model. Although GhostXL-v1.0 wasn't as good as I had imagined, it was released because of its unique style and passed its own level in every way. Before using GhostXL, I recommend reading the instructions below, thank you. Abou* ***ng GhostXL-v1.0 1. Actual measurements can directly generate 1024-1536,1024-1536. Highly detailed images do not require high-definition restoration or refiner. 2. xx art style. You can fine-tune the style through xx art style. The following are the results obtained by adding the same pattern, the same seed, and different art styles. (PS: If you want to get effects similar to GhostMix, I personally recommend Realistic/Realistic anime art style) <img src="file:///C:\Users\BRENDA~1\AppData\Local\Temp\ksohtml23864\wps1.png" alt="" data-href="" style=""/> <img src=” https://liblibai-online.liblib.cloud/img/59e9d6328afd833dd67b58f46db1ee2f/2bf56eb03451e2028574fa12f1a6336e5e38b9d07613d96389125c4003ee7964.jpg?x-oss-process=image/resize,w_1290,m_lfi t/format, webp” alt= "” data-href=” https://liblibai-online.liblib.cloud/img/59e9d6328afd833dd67b58f46db1ee2f/2bf56eb03451e2028574fa12f1a6336e5e38b9d07613d96389125c4003ee7964.jpg?x-oss-process=im age/resize, w_1290, m_lfit/format, webp” style= "” /> 3. Personal impression: The compatibility is worse than GhostMix, the picture quality is slightly weaker, and the image yield is slightly higher. (The comparison here is all GhostMix that has been repaired in HD) But since there are no models in this style on the market, I released GhostXL. The advantage is that GhostXL can run straight out without HD restoration. 4. In terms of negative reminders: The sample image uses negativeXL_D embedding, and the effect is quite good. If you need to reproduce it, please download the embedding yourself. Analysis of the advantages and disadvantages of GhostXL-v1.0 As mentioned in my previous post, I mainly considered the model: 1. Compatibility 2. Picture Quality 3. Good Picture Ratio. First of all, in terms of compatibility, GhostXL is worse than GhostMix. GhostXL-v1.0 is a 2D-2.5D painting model and has poor compatibility with 3D painting styles. GhostXL is based on the Animagine XL V3 finetune and then made by integrating other models (the finetune model accounts for 70% of the total model, 30% of the other models). The 1.0 version of GhostXL is still only a 2D-2.5D model. I will continue to develop and see if I can improve GhostXL's style compatibility. Thanks to the compatibility of Animagine XL V3, GhostXL has a good response to Masterpiece, Best Quality, and anime characters. However, some stylized promtps responses are poor. Then there's the picture quality. I personally feel that the overall picture quality is OK. After all, GhostXL is a direct 1024-1536 generation (one-stage generation), not a two-stage generation with high-definition restoration after 512-768 like GhostMix. In the case of one-stage generation, you can have this image quality, and I'm personally quite satisfied. Finally, the good image rate. I personally feel that the good image rate is still higher than GhostMix. However, there isn't much difference between the same prompts with GhostMix; you need to adjust the prompts to make the difference between images under the same theme. In order to facilitate user feedback and communication and cooperation, if interested, please add me the details of the V:ghostinshell10 model <ul><li>Model Author: GhostInshell Model Type: Diffusion-based text-to-image</li> <li>generative model</li> <li> License: Fair AI Public License 1.0-SD</li> <li style="text-align: start;"><li>Fine Tuned to Model</li>: Animagine XL 3.0 Config </li><li>of Finetuning: Ghostxl_finetune.toml OthermergeModel: Leosam's Hello World XL by Big Rabbit</li> <li> DreamShaper XL</li></ul>

Files with SHA256: ee1fdc86f76e71ca223d671da5243c0103e9f2e3d8e80a49c9bdf24b991ba6c0