Recommended weight：1
v2 note:
I added a lot of training material related to depth of field, while keeping the goal of proportion redistribution in mind during training. Additionally, the purpose was to prevent characters from being too far and becoming blurry.
Two epochs that were closer to the target were merged together, and the results were pretty good, especially regarding reflection effects.
Since the art style has changed significantly in this version, I've categorized it under the "style" classification.
v1 note:
A personal experimental model with the goal of redistributing the proportions between characters and scenes, while enhancing the lighting and details of the scene, such as reflections.
The result is decent, but when characters are zoomed out, facial blurring may still occur. It is recommended to use hires fix or extensions like adetailer to address this issue.
The weight can be adjusted according to personal preference.
Feel free to leave feedback in the comments below, or share images generated using this model!
______________________________________________________________________________________________
建議權重：1
v2 note:
新增許多景深相關的訓練素材，同時保持比例重分配的目標概念來進行訓練，另外也有讓人物佔位不要太遠導致模糊的訓練目的,
採用兩個比較接近目標的epoch進行融合，效果還蠻不錯的，特別是關於反射相關的效果,
另外該版本畫風變動較大，因此改成style分類,
v1 note:
個人實驗用的模型，目標是將人物和場景的比例重新分配，同時強化場景的光影和細節，例如反射等,效果還行，只是人物一旦拉遠仍會有面部模糊的問題，建議高清修復或是使用adetailer之類的擴充插件,
可照個人目標調整權重
歡迎底下留言反饋，或是用這模型產出圖片返圖!
描述:
训练词语:
名称: fos_v2_ab.safetensors
大小 (KB): 166541
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
