coreml-YamersRealistic-v4_SDXL_8-bit

Core ML Converted SDXL Model:

• This model was converted to Core ML for use on Apple Silicon devices. Conversion instructions can be found here.
• Provide the model to an app such as Mochi Diffusion Github / Discord to generate images.
• original version is only compatible with CPU & GPU option.
• Resolution is the SDXL default of 1024x1024 UNLESS NOTED OTHERWISE.
• This model requires macOS 14.0 or later to run properly.
• This model is quantized to 8-bits.
• This model was converted with a vae-encoder for use with image2image.
• 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.
• This model does not include a safety checker (for NSFW content).
• This model can not be used with ControlNet.

<br>

YammersRealistic-v4_SDXL_8-bit:

Source(s): CivitAI<br>

This is an SDXL base model converted and quantized to 8-bits.<br><br>

<br><br>

Yamer's Realistic v4 VAE NSFW & SFW

Support me on Twitter: @YamerOfficial

Discord: yamer_ai

Yamer's Realistic is a model focused on realism and good quality, this model is not photorealistic nor it tries to be one, the main focus of this model is to be able to create realistic enough images, the best use with this checkpoint is with full body images, close-ups, realistic images and beautiful sceneries, for example futuristic cities or insides of a home. Give this model a chance and let your creativity guide you and the perception you have on the world!

This model is not nsfw focused but it can crate nsfw content, it is preferrable to use a LoRA for specific nsfw content, I also recommended to use words such as: Realistic, Ultra realistic, Photorealistic etc..., to give the resulting image an even better and more accurate realistic look to it.

If you want to explore fields that are not realistic focused I highly suggest to use my LoRA for abstract work and a better ethereal look: Yamer's Style.

No refiner is needed for this model to produce good quality results.

A detailer is not required to achieve a realistic looking image.

The main style of this model is realistic but that can change with a specific prompt that aims at an anime look for example.

Thank you for using my models and writing a review, all forms of support are appreciated, it takes me a lot of time to produce this kind of checkpoint every time, thank you again! ❤️

Version 4 + VAE comes with the SDXL 1.0 VAE already baked in.

Recommended settings:

• Image resolution: 1024x1024 (standard SDXL 1.0 base resolution)
• Steps: 30 to 150 (Lower steps may produce image artifacts and other unwanted results)
• Restore Face: Not needed to generate high quality images (Not used in the showcase)
• Clip skip: 1 and 2 (Works good with both settings)
• CFG: 7 to 25