Wolverine (Logan) X-Men Evolution [LoRA]版本v1.1 Suit 1+L Noob-E v1.0 (ID: 1212942)

LoRAs is trained on screenshots of X-Men Evolution (2000-2003) taken from the website fancaps.net

v1.0 BETA - (Only Logan) without Wolverine suits! The model can sometimes create black borders, this will be fixed in the full version! LoRA is trained on 230 screenshots

v1.0 - There are no more black borders, a little more sharpness and fewer JPEG artifacts. There are Wolverine first and second suit. LoRA is trained on 480 screenshots

v1.1 - The main difference from v1.0 is that I split the data and created several LoRAs, the previous version was something like all in one, but it seems to me it was not a very good idea. I changed the captions a bit and made almost no changes to the images.

Logan - (Only Logan) Only Logan without the Wolverine suits

Logan Lite - (Only Logan) Lite - Only Logan in dark blue t-shirt and brown jacket without cowboy hat (The dataset has been reduced by almost half)

Suit 1 - (First Suit) Only Wolverine in first suit, LoRA was not trained on images of Wolverine without a mask!

Suit 1+U - (First Suit+Unmasked) Only Wolverine in first suit + Images of Wolverine without a mask

Suit 1+L - (First Suit+Logan) Only Wolverine in first suit + Images of Wolverine without mask and images of Logan to enhance the creation of Wolverine without Mask

Suit 2 - (Second Suit) Only Wolverine in the second suit, without images of Logan

Suit 2+L - (Second Suit+Logan) Only Wolverine in the second suit and images of Logan

Preview generated without using ADetailer & Hires. fix

Recommended LoRA Strength (Weight): 0.8-0.9 - 0.7-0.8 BETA

描述:

Repeats: 1

Epoch: 7

Image Count: 322

Network Rank (Dimension): 16

Network Alpha: 8

Batch size: 1

Clip skip: 1

Training steps: 2254

Resolution: 1024x768

Train Resolution: 1024,1024

Model: noobaiXLNAIXL_epsilonPred10Version

LR Scheduler: cosine_with_restarts

Optimizer: AdamW8bit

Learning rate: 0,0005

Text Encoder learning rate: 0,00005

Unet learning rate: 0,0005

Captioning: WD14 Captioning (SmilingWolf/wd-v1-4-moat-tagger-v2) + Edited

Recommended LoRA Strength (Weight): 0.8-0.9

训练词语: wolvevo, bodysuit, orange bodysuit, black bodysuit, two-tone bodysuit, short sleeves, mask, covered eyes, white eyes, no pupils, belt,,logan, wolvevo, hood down, bodysuit, orange bodysuit, black bodysuit, two-tone bodysuit, short sleeves, orange gloves, orange footwear, belt,