
Use Art by Nekuzx
as trigger. Trained on couple hundred images of Nekuzx. My first time creating a LoRA and results were random at the start of testing. Managed to make it more stable as shown in examples by adding score_7_up
to the prompt to inherit more quality from pony v6 xl but not higher scores because they seem to rewrite the intended style.
was using this settings for examples:
Steps: 30, Sampler: Euler a, Schedule type: Automatic, CFG scale: 7
Going to make this more stable and controllable once I'll collect enough buzz and rewrite dataset by myself.
描述:
First time making a LoRA. Not perfect but working.
训练词语: Art by Nekuzx
名称: 662400_training_data.zip
大小 (KB): 297011
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: Nekuzx_artstyle.safetensors
大小 (KB): 223135
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success