💚 for Updates - I would love to see your reviews and images of this model, feel free to post them below! Ambience is currently an experimental test model that aims to replicate the creativity and styling of <a target="_blank" rel="ugc" href="https://civarchive.com/models/115094/maverick">Maverick</a> using the SDXL 1.0 base. I'm currently unable to test and operate these models as I would normally on Automatic1111 and have recently tried out Comfy UI to be able to run XL base models.💡 This model is currently an experimental project<ul><li>expected to change a good amount from version to version.</li><li>testing is currently slow, limited, &amp; inefficient.</li><li>if possible with my gpu/vram, I would like to look into training some simple elements to merge in. I may look into Google Colab as an option as well.</li><li>I'm hoping to get XL based models functioning on A1111 for better testing and familiar workflow.</li><li>current aim is to merge in more raw elements, sharpen things a bit with clarity to get rid of the rough/feathery aesthetic with the XL base model currently, improving backgrounds/scenery, clean up any detail inaccuracies where I can.</li></ul><hr /><h3 id="heading-1841">Recommended Use</h3>Unknown for the most part at the moment. Below are a few example workflows using ComfyUI. I have been using the SDXL base refiner, but this may not even be needed. You can also try using a refine element with the model itself, a SD 1.5 base model, or a similar XL model. I have not been able to try any hires fix options and haven't done much upscaling where I'm able to try denoising parameters etc. Currently I've been using ~30 steps, 7 scale, DPM++ SDE Karras, Euler a, or a mixture of the two split between base model/refinerYou can try other samplers as well. DPM++ 2M Karras and Euler seemed to be decent too. I'm not really sure how a clip skip parameter is replicated in Comfy UI at the moment. CFG Scale I have been fine with at 7, you can try higher or lower and try splitting different variables between base model/refiner if used. Step count I have kept at mostly 30, 20-50 probably seem good. Feel free to see what works best with other parameters.Negative embeddings are not needed. I have used some here and there and they seem to work fine. You may find positives or negatives with them depending on use.<hr /><h3 id="heading-1841">Workflow Examples</h3>The preview images should have workflows/metadata built in and should be viewable by dragging into your interface (Comfy ui, A1111, etc.). Having just tried out Comfy ui myself recently, you may find you have a more adequate workflow(s).<ul><li><a target="_blank" rel="ugc" href="https://rentry.co/pdr28">Example 1</a> (slightly rearranged, adds additional quality tags within the workflow - you can change these to best match your styling, initially via KidCharles)</li><li><a target="_blank" rel="ugc" href="https://rentry.co/szfab">Example 2</a> (workflow from the SDXL page, via StabilityAI or Civitai)</li><li><a target="_blank" rel="ugc" href="https://rentry.co/batn3">Example 3</a> (simple SDXL workflow, via Comfy ui creator? unknown source)</li></ul>There are also many great Comfy ui workflows listed on Civitai, plugins, etc. Such as ones geared toward <a target="_blank" rel="ugc" href="https://civarchive.com/models/117996/sdxl-config-comfyui-fast-generation-4gb-vram-refiner">4GB vram cards</a>. Try filtering to the 'other' category and/or searching for them with SDXL/workflow keyword(s). <hr />Disclaimer<pre><code>This model(s) may output NSFW content unintentionally depending on parameters used. Make sure you tailor your prompts accordingly. For example "nsfw" in the negative prompt.

The purpose of sharing this model is not to showcase obscene material in a public forum. The use of this learning model is entirely at the discretion of the user, and they have the freedom to choose whether or not to create SFW or NSFW content. The decision of whether to engage with SFW or NSFW content lies with the user and their own personal preferences. The AI model(s) do not contain explicit visual content that can be accessed easily.</code></pre>Note: this is a general disclaimer, I have not experienced any nsfw content popping up with SDXL based models thus far.<hr /><h3 id="heading-55">Resources</h3>This model merge (a1) used the following resources - all of these are great stand-alone, props to each creator.<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/117359/mbbxl-alpha">MBBXL Alpha_v0.1</a> (via <a target="_blank" rel="ugc" href="https://civarchive.com/user/morbuto/models">morbuto</a>) note: 'no sharing merges' is now listed on v1.0RC and was not present on the model page with v0.1, I do not plan to use any versions of this going forward.</li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/112902/dreamshaper-xl10">DreamShaper XL 1.0_alpha 2 </a>(via <a target="_blank" rel="ugc" href="https://civarchive.com/user/Lykon/models">Lykon</a>)</li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/117041">WowifierXL</a> (via <a target="_blank" rel="ugc" href="https://civarchive.com/user/mooncryptowow/models">mooncryptowow</a>)</li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/117518?modelVersionId=128996">Headshot - SDXL_v1</a> (via <a target="_blank" rel="ugc" href="https://civarchive.com/user/ntc/models">ntc</a>)</li></ul><hr /><h3 id="heading-157">License &amp; Use</h3><a target="_blank" rel="ugc" href="https://github.com/Stability-AI/generative-models/blob/main/model_licenses/LICENSE-SDXL1.0">CreativeML Open RAIL++-M</a>Use Restrictions<pre><code>You agree not to use the Model or Derivatives of the Model:
- In any way that violates any applicable national, federal, state, local or international law or regulation
- For the purpose of exploiting, harming or attempting to exploit or harm minors in any way
- To generate or disseminate verifiably false information and/or content with the purpose of harming others
- To generate or disseminate personal identifiable information that can be used to harm an individual
- To defame, disparage or otherwise harass others
- For fully automated decision making that adversely impacts an individual’s legal rights or otherwise creates or modifies a binding, enforceable obligation
- For any use intended to or which has the effect of discriminating against or harming individuals or groups based on online or offline social behavior or known or predicted personal or personality characteristics
- To exploit any of the vulnerabilities of a specific group of persons based on their age, social, physical or mental characteristics, in order to materially distort the behavior of a person pertaining to that group in a manner that causes or is likely to cause that person or another person physical or psychological harm
- For any use intended to or which has the effect of discriminating against individuals or groups based on legally protected characteristics or categories
- To provide medical advice and medical results interpretation
- To generate or disseminate information for the purpose to be used for administration of justice, law enforcement, immigration or asylum processes, such as predicting an individual will commit fraud/crime commitment (e.g. by text profiling, drawing causal relationships between assertions made in documents, indiscriminate and arbitrarily-targeted use).</code></pre>Terms of use<pre><code>- You are solely responsible for any legal liability resulting from unethical use of this model(s)
- If you use any of these models for merging, please state what steps you took to do so and clearly indicate where modifications have been made.</code></pre>Note<pre><code>If you see any conflicts or corrections to be made, please let me know.</code></pre>

Ambience - SDXL

Release - Alpha 1Experimental refinements with merging various SDXL models.Note: this version/model is not finetuned to maximum detail accuracy.

Files with SHA256: 2c507544192c6668c008aaba816f4bea7e0c9b09d8f818618873b6750a8faf6c