
The updated version 2 had a complete rework of the Dataset. SDXL and Flux versions was created with it as well.
I am quite happy with how this Lora turned out! Enjoy!
The Lora is trained on MidJourney Images that use my "Personalised Style" +Oil Pastels
Works without the trigger words ArsMJStyle, Oil Pastels
Have some small nice aesthetic enchancment effects in the 0.2 - 0.8 range.
Gives intended results in the 0.8 - 1.2 range .
I love the results at 1.5 .
描述:
The Lora is trained on MidJourney Images that use my "Personalised Style" + Pastels
Works without the trigger words ArsMJStyle, Pastels
Try lowering CFG if everything is too "melty"
Have some small nice aesthetic enchancment effects in the 0.2 - 0.4 range.
For general purpose use 0.4 - 0.8
For Still Life and Landscapes you can use 0.8 - 1.2 . This was the bulk of the dataset (only ~20% portraits).
Depends on the actual Pony model you will use, but does not produce good results above 1.2
训练词语: ArsMJStyle, Pastels
名称: 620866_training_data.zip
大小 (KB): 82317
类型: Training Data
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
名称: ArsMJStylePony_-_Pastels.safetensors
大小 (KB): 223102
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success