Sign In

トパーズ - Topaz LoRa / Honkai: Star Rail - 崩壊スターレイル

136
975
11.1k
15
Verified:
SafeTensor
Type
LoRA
Stats
975
11,141
Reviews
Published
Oct 20, 2023
Base Model
SD 1.5
Training
Steps: 1,999
Epochs: 10
Usage Tips
Clip Skip: 2
Trigger Words
TopazV4
Hash
AutoV2
CECEFE67D7
KXRO's Avatar
KXRO

N°50 Release, トパーズ - Topaz from Honkai: Star Rail - 崩壊スターレイル.

Put everything in Filter for seeing all images!

Click on Show more for Weight & Trigger recommendation, Also click on About This Version at Right Corner for more Info.

I wanted to create her in the past, Quality is Ok/Good & Eyes is pratically right in colors + Flexibility Ok.

About this Lora:

  • This Lora is trained on 112+ Images at the Max Quality Possible, (Majority Of Orginal Art 4K + Official Art 4K & Some Like 4-5 GameScreenCaps), I sorted everything manually, Only picked the best possible. (SFW/NFSW+) - All Artwork used in Datasets are all updated as of today. (19/10/2023).

  • I Removed all the textmark from Datasets + Upscaled some Artwork with x4-AnimeSharp for a Better Output Quality.

  • NAI Used, Work with Majority of the Anime Model's.

  • Flexibility is Not Bad/Good. (See Examples Images). - (0.6-0.8 Weight, you can even go with Weight 1 in some cases).

  • Good Precision on Official Default Skin & Can do Also Other Clothes that don't exist. (Tho only Downside is the Knee Boots, but you can get it pratically right too).

  • Can do Original Art.

  • Increased size 74Mb + (With Optimal Settings), Like this Multicolored Eyes are a little more Consistent + Precise.

  • Work with other Lora like Add Detail & Other. (+ ADetailer i recommend).

  • Can do in Realistic Cosplay.

About Trigger Words & Prompts: TopazV4 (For Trigger) & for prompts just refer to examples image.

Weight recommended: 0.6-1. - Work perfectly at Weight 1, Best result: 0.8-1 (It's Varying per Model's, Majority of time 0.8 Weight is Better/Flexibility).

Negative Embedding used: EasyNegative, BadPrompt2 and sometime negative_hand-neg.

Leave a Review & Share you're image in comments if you appreciate this Work :)