---
license: creativeml-openrail-m
tags:
- coreml
- stable-diffusion
- text-to-image
- not-for-all-eyes
---
# Core ML Converted Model:
- This model was converted to [Core ML for use on Apple Silicon devices](https://github.com/apple/ml-stable-diffusion). Conversion instructions can be found [here](https://github.com/godly-devotion/MochiDiffusion/wiki/How-to-convert-ckpt-or-safetensors-files-to-Core-ML).
- Provide the model to an app such as **Mochi Diffusion** [Github](https://github.com/godly-devotion/MochiDiffusion) / [Discord](https://discord.gg/x2kartzxGv) to generate images.
- `split_einsum` version is compatible with all compute unit options including Neural Engine.
- `original` version is only compatible with `CPU & GPU` option.
- Custom resolution versions are tagged accordingly.
- This model was converted with a `vae-encoder` for use with `image2image`.
- This model is `fp16`.
- Descriptions are posted as-is from original model source.
- Not all features and/or results may be available in `CoreML` format.
- This model does not have the [unet split into chunks](https://github.com/apple/ml-stable-diffusion#-converting-models-to-core-ml).
- This model does not include a `safety checker` (for NSFW content).
# QGO-10b:
Source(s): [CivitAI](https://civitai.com/models/4188/qgo-10b)
The biggest difference here has not been the merge, but just cutting down on negative prompt. This can have amazing results for realism, though you may run into things that are a little too real. Be warned ;)
It uses RPGv4 instead of v3. It also involves Latex to get some of the bondage model back in as well..
There may also be traces of the Middle Finger and Gun2Head Pose LORAs in the examples.
Advisable to use Hires.fix with the following (or similar) settings:
Upscaler: ESRGAN_4x (NMKD superscale can be a bit sharper, which is nice for smaller upscales)
Upscale by: 1.1~2.0 (whatever suits your purpose)
Denoising strength: 0.3 (or 0.5~0.7 if you don't mind changes from the base image)
You can of course generate low resolution versions first, and pick out the ones you enter into the upscale process (saves a lot of time). I really takes the amount of detail, especially of the face and eyes, to the next level.
This model aims for photorealism at higher resolutions, and large variation in poses, settings and genres. (And having a lot of fun filling half the prompt with wildcards).
I will still be tweaking it in the near future, to eliminate unwanted outputs and/or increase possibilities. Let me know if you run into anything that seems off.
Shout-out to the Unstable Diffusion Discord, where all the cool people share their gems in #photorealistic
