
❗Update 2❗ Leave a like if you like it.
better upscaling
❗Update 1❗
pictures made with dev-Q8_0.gguf model.
It shouldn’t appear as "rooty" anymore. The quality has also improved overall. While I still wouldn’t recommend using strength:1 for highly detailed LoRAs. It can work well in some cases, but sometimes there’s just too much detail. I still recommend using a strength of 0.5 to 0.7. I’ll use the rose ship as an example to show the results before and after the update, both at strength:1. The slight blurriness is from resizing the image.?
Here’s my very own Fantasy Enhancer! I know there are a few out there but I wanted to try it myself.
Try using it with my New Fantasy Core Flux LoRA. Check it out here: New Fantasy Core Flux on Civitai. Enjoy experimenting!
Recommended strength of 0.5 if you try it with other LoRA´s.
Now that I know how to do it, V2 will get a massive quality upgrade... but flux training is expensive on civitai and I would like to try other parameters
描述:
训练词语:
名称: FinishMyFantasy .safetensors
大小 (KB): 522341
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
名称: FinishMyFantasy_update 2 FLUX.safetensors
大小 (KB): 18806
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
名称: neu5.safetensors
大小 (KB): 18806
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success