feat: Add Diffusers

[mudler]

Description

This PR fixes #589 and adds a preliminar support for diffusers.

For example, to use AnimagineXL (but it's not limited to only that one!) create a model YAML config file like this (note this is forcing CPU usage and it's quite slow ~15m for an image - for GPU set cuda to true and f16 to true):

cat models/anime.yaml    
name: anime
parameters:
  model: Linaqruf/animagine-xl
backend: diffusers

# Force CPU usage - set to true for GPU
f16: false
diffusers:
  pipeline_type: StableDiffusionXLPipeline
  cuda: false # Enable for GPU usage (CUDA)
  scheduler_type: EulerAncestralDiscreteScheduler

From the API call:

curl http://localhost:8080/v1/images/generations -H "Content-Type: application/json" -d '{
            "prompt": "positive|negative", "model": "anime", "step": 51,
            "size": "256x256" 
          }'

Example

curl http://localhost:8080/v1/images/generations -H "Content-Type: application/json" -d '{
            "prompt": "face focus, cute, masterpiece, best quality, 1girl, green hair, sweater, looking at viewer, upper body, beanie, outdoors, night, turtleneck|lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry", "model": "anime", "step": 51,
            "size": "1024x1024" 
          }'

Notes:

• the model is not officially supported by the OpenAI client - you can specify the model in the Bearer token when using the official OpenAI client
• step can be configured as well in the YAML config file to a fixed value

Signed commits

• Yes, I signed my commits.