The last three images are the original images of my training
Ben Lora:
Base mold:
leosamsMoonfilm filmGrain20.safetensors [fe54b5d04d] 【Cinematic realism】
revAnimated 122.safetensors (4199bcdd14) [two-dimensional style]
VAE:
vae-ft-ema-560000-ema-pruned.ckpt 【Filter bright】
ldaenxiwasueJo [filter cold light]
If you download and install the above models and VAE locally, you can find them on the mile cloth or C station, of course, you can also use the base mold and filter with a similar effect, as long as the effect of the picture is satisfactory
Here I suggest the next three SD plugins (those that have been downloaded, you can skip them)
sdweb-easy-prompt-selector (adjusts the weight of lora) for example: <lora:XM:0.7>
ultimate-upscale-for-automatic1111
adetailer (face and detail restoration)
ultimate-upscale-for-automatic1111 and HD repair recommend that the graphics card be at least 3060 or higher, otherwise it is difficult to run
This LoRa uses:
High resolution fix without opening