
1. This Lora is of the pony type. When using it, make sure to select a large model of the pony type; otherwise, you won't achieve the expected results.
2. The best results are achieved when the model weight is set to 0.8.
3. If the model severely breaks down, it could be due to the following reasons: mismatched large model, no input prompt, model weight too high, or trigger word conflicts (as some words were used as special trigger words during training, using multiple words together may result in broken images).
4. Different forms of the character can be activated through special trigger words, such as "white hood, gloves," "seaside, shirt, blue underwear," "jacket," etc., which can trigger different character forms. You can refer to the sample images for specific examples.
5. The eyes of the model are most prone to breaking down. You can use other Lora models to assist in fixing this.
6. Thank you all for your support, updates will be released periodically.
描述:
The model has been improved, and the current output results are closer to the original art style.
Starter words:gabu,bear boy,furry,green eyes,brown fur,score_9_up,score_8_up,source_furry,male focus,plump,<lora:Gabu-000020:0.8>,(perfectly detailed, detailed face, detailed eyes, beautiful eyes)
Aside from the starter words, if no additional input is provided, the generated image will default to a close-up photo similar to an emoji. If you want to create various fun emojis, you can directly add an emoji hint after the starter words. Adding more hint words on this basis will enrich the content.
If you encounter any issues with partial distortion, you can reduce the prompt or use the ADetailed plugin to resolve it.
训练词语: gabu,furry,green eyes,bear boy
名称: Gabu-000020.safetensors
大小 (KB): 223100
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success