Model Introduction:
This is a two headed Q version of anime style anime vertical drawing Lora!
Please refer to the example picture for specific consumption methods!!
Recommended parameters:
Large model: anything_inkBase
VAE: Automatic
CLIP skip: 1
Recommended positive tags: chibi, white background, full body,
Recommended negative tags: logo, text, badhandv4, EasyNegative, ng_deepnegative_v1_75t, bad_prompt Negative Embedding_v2.0, bad image repair veryBadImageNegative_veryBadImageNegative_v1.3, negative hand neg, poor drawn face, extra limb, missing limb, disconnected limbs, malformed hands, ugly,
Trigger word: Ertouqchibi,
Lora weight: 0.4-0.5 (this range is good)
Iteration steps: 30
Sampling recommendation: DPM++2M karras
Resolution: 512 * 512
Encountered a good-looking one during generation, requires high-resolution repair! Enlarge the algorithm by selecting R-ESRGAN 4x+and redraw with an amplitude of 0.3!
Finally:
Need to draw cards, wish everyone a happy time, welcome to test and return pictures!!